Jorge Enrique García-Farieta

Context. Halo bias links the statistical properties of the spatial distribution of dark matter (DM) halos to those of the underlying DM field, providing insights into clustering properties in both general relativity (GR) and modified-gravity such as f(R) models. While the primary halo mass-dependent bias has been studied in detailed, the secondary bias, which accounts for the additional dependencies on other internal halo properties, can offer a sensitive probe for testing gravity beyond the Λ-cold dark matter (ΛCDM) model. Aims. The aim of this study is to quantify any potential deviations between ΛCDM and f(R) gravity models in halo clustering, at both the primary and secondary level, as well as in the distributions of halo properties in the cosmic web. Methods. Using N-body simulations of f(R) gravity models, we assessed the scaling relations and the primary and secondary bias signals of halo populations on the basis of a halo-by-halo estimator of large-scale effective bias. Our analysis was performed using halo number density as the independent variable. Results. The relative difference in the effective bias between the f(R) models and ΛCDM is sensitive, albeit slightly, to the power index of modified gravity. The largest deviations from GR are measured for low-mass halos, where the average bias at a fixed number density decreases by up to 5% for fixed scaling indices. We also show that the scaling relations for some environmental properties, including neighbour statistics, Mach number, and local overdensity, exhibit small but non-negligible deviations (3–5%) from GR for a wide range of number densities. Our results suggest that the properties of halos in sheets and voids show the largest departures from GR (> 10% in some cases). In terms of secondary bias, we do not find any statistically significant deviations with respect to ΛCDM for any of the properties explored in this work.

Context. We assess the effectiveness of a non-parametric bias model in generating mock halo catalogues for modified gravity (MG) cosmologies, relying on the distribution of dark matter from either MG or Λ cold dark matter (ΛCDM) simulations. Aims. We aim to generate halo catalogues that effectively capture the distinct impact of MG, ensuring high accuracy in both two- and three-point statistics for a comprehensive analysis of large-scale structures. We investigated the inclusion of MG in non-local bias to directly map the tracers onto ΛCDM fields, which would significantly reduce computational costs. Methods. We employed the bias assignment method (BAM) to model halo distribution statistics by leveraging seven high-resolution COLA simulations of MG cosmologies. Taking cosmic-web dependences into account when learning the bias relations, we designed two experiments to map the MG effects: one utilising the consistent MG density fields and the other employing the benchmark ΛCDM density field. Results. BAM generates MG halo catalogues from both calibration experiments with excellent summary statistics, achieving a ~1% accuracy in the power spectrum across a wide range of k modes, with minimal differences well below 10% for modes subject to cosmic variance, particularly below k < 0.07 h Mpc⁻¹. The reduced bispectrum remains consistent with the reference catalogues within 10% for the studied configuration. Our results demonstrate that a non-linear and non-local bias description can model the effects of MG starting from a ΛCDM field.

Accurate modeling of galaxy distributions is paramount for cosmological analysis using galaxy redshift surveys. However, this endeavor is often hindered by the computational complexity of resolving the dark matter halos that host these galaxies. To address this challenge, we propose the development of effective assembly bias models down to small scales, i.e., going beyond the local density dependence capturing non-local cosmic evolution. We introduce a hierarchical cosmic web classification that indirectly captures up to third-order long- and short-range non-local bias terms. This classification system also enables us to maintain positive definite parametric bias expansions. Specifically, we subdivide the traditional cosmic web classification, which is based on the eigenvalues of the tidal field tensor, with an additional classification based on the Hessian matrix of the negative density contrast. We obtain the large-scale dark matter field on a mesh with ~3.9 h ^-1 Mpc cell side resolution through Augmented Lagrangian Perturbation Theory. To assess the effectiveness of our model, we conduct tests using a reference halo catalogue extracted from the UNIT project simulation, which was run within a cubical volume of 1 h ^-1 Gpc side. The resulting mock halo catalogs, generated through our approach, exhibit a high level of accuracy in terms of the one-, two- and three-point statistics. They reproduce the reference power-spectrum within better than 2 percent accuracy up to wavenumbers k ~ 0.8 h Mpc^-1 and provide accurate bispectra within the scales that are crucial for cosmological analysis. This effective bias approach provides a forward model appropriate for field-level cosmological inference and holds significant potential for facilitating cosmological analysis of galaxy redshift surveys, particularly in the context of projects such as DESI, EUCLID, and LSST.

Context. The new generation of galaxy surveys will provide unprecedented data that will allow us to test gravity deviations at cosmological scales at a much higher precision than could be achieved previously. A robust cosmological analysis of the large-scale structure demands exploiting the nonlinear information encoded in the cosmic web. Machine-learning techniques provide these tools, but no a priori assessment of the uncertainties. Aims. We extract cosmological parameters from modified gravity (MG) simulations through deep neural networks that include uncertainty estimations. Methods. We implemented Bayesian neural networks (BNNs) with an enriched approximate posterior distribution considering two cases: the first case with a single Bayesian last layer (BLL), and the other case with Bayesian layers at all levels (FullB). We trained both BNNs with real-space density fields and power spectra from a suite of 2000 dark matter-only particle-mesh N-body simulations including MG models relying on MG-PICOLA, covering 256 h⁻¹ Mpc side cubical volumes with 128³ particles. Results. BNNs excel in accurately predicting parameters for Ω_m and σ₈ and their respective correlation with the MG parameter. Furthermore, we find that BNNs yield well-calibrated uncertainty estimates that overcome the over- and under-estimation issues in traditional neural networks. The MG parameter leads to a significant degeneracy, and σ₈ might be one possible explanation of the poor MG predictions. Ignoring MG, we obtain a deviation of the relative errors in Ω_m and σ₈ by 30% at least. Moreover, we report consistent results from the density field and power spectrum analysis and comparable results between BLL and FullB experiments. This halved the computing time. This work contributes to preparing the path for extracting cosmological parameters from complete small cosmic volumes towards the highly nonlinear regime.

Earth's magnetosphere, beyond protecting the ozone layer, is a natural phenomena which allows to study the interaction between charged particles from solar activity and electromagnetic fields. In this paper we studied trajectories of charged particles interacting with a constant dipole magnetic field as first approach of the Earth's magnetosphere using different initial conditions. As a result of this interaction there is a formation of well defined radiation regions by a confinement of charged particles around the lines of the magnetic field. These regions, called Van Allen radiation belts, are described by classical electrodynamics and appear naturally in the numerical modeling done in this work.