Marta Spinelli

Context. Foreground cleaning is one of the main challenges for H I 21 cm intensity mapping, here considered over the redshift range z ≃ 0.09 − 0.58, particularly in the presence of a chromatic instrumental beam that couples angular foreground structure into the frequency direction. This problem is further complicated by incomplete sky coverage, as expected for realistic Galactic masks and survey footprints. Aims. We introduce JESSNet, a beam-aware and mask-compatible foreground-cleaning framework designed to preserve the statistical properties of the H I signal in single-dish intensity mapping observations affected by frequency-dependent beam distortions and masked sky regions. Methods. JESSNet advances the joint deconvolution and blind source-separation framework of DecGMCA and SDecGMCA through three methodological extensions. First, while the original formulation uses a single effective mixing model over the angular range being cleaned, JESSNet operates in angular windows, with an effective mixing matrix and a channel selection adapted to each scale. Second, JESSNet replaces the fixed starlet thresholding step with a learned spherical-wavelet “learnlet” sparse regularization operator. Third, whereas the original spherical implementation assumes full-sky coverage, JESSNet constrains the source estimates to the observed sky region, making the method applicable to Galactic masks and survey footprints. We tested the method on simulated SKA-MID-like observations covering 900 − 1300 MHz, including H I emission, Galactic and extragalactic foregrounds, thermal noise, and an oscillating in frequency MeerKAT-inspired beam model. Results. The multiscale JESSNet reconstruction accurately recovers the angular and frequency power spectra of the input H I signal. Compared to applying JESSNet over a single angular range, the multiscale formulation significantly improves the reconstruction at low and intermediate angular scales and reduces residual beam-induced spectral features. The method also remains effective for a synthetic analysis footprint, preserving the H I power spectra under incomplete sky coverage. Conclusions. JESSNet provides a beam-aware, multiscale foreground-cleaning framework for H I intensity mapping. The code used to produce the analysis and results has been publicly released, allowing the community to reproduce the pipeline and adapt it to future simulations and real observations.

The redshifted 21cm line of Neutral Hydrogen (HI), detected by radio telescopes can probe the large-scale structure of the post-reionization Universe via Intensity Mapping techniques. This makes HI one of the most promising observables to transform our understanding of dark matter and dark energy. We present here what we hope to achieve once the full SKAO is operational. We also discuss the recent development on HI intensity mapping with the SKAO precursor MeerKAT and a key challenge in these types of experiments: the subtraction of the bright foregrounds. The subtraction can be achieved with component separation techniques, but the quality of the results needs to be carefully checked against simulations. Fortunately, the cross-correlation of the HI signal with galaxy survey data which trace the same underlying dark matter distribution allows a strong reduction of systematics down to the detection of the HI signal in current MeerKAT data.

The redshifted 21 cm signal from the post-reionization epoch is highly non-Gaussian; thus, higher-order statistics, such as the bispectrum, are required to extract this non-Gaussian information. However, high signal-to-noise ratio (SNR) detection of the 21 cm auto-bispectrum will be hindered by the presence of residual systematics. Cross-correlating the 21 cm signal with galaxies offers a promising path to suppress this uncertainty from residual systematics and potentially increase the SNR. We present a comprehensive analysis of the H I -galaxy cross-bispectrum using the predictions of theoretical galaxy evolution models defined on large cosmological volumes. Our analysis includes the cross-bispectrum for different triangle sizes and shapes, as well as for different combinations of the H I and galaxy fields. We forecast the detectability of the 21 cm-galaxy cross-bispectrum at redshift z ≈ 1 with the Euclid-like galaxy survey and SKA-Mid observations in both the interferometric and single-dish modes of the survey. We find that the 21 cm-galaxy cross-bispectrum shows enhanced detectability compared to the 21 cm auto-bispectrum for all unique triangles in the interferometric mode of observations. We forecast a 10 σ detection of the cross-bispectrum for squeezed-limit triangles and a 100 σ detection for all shapes combined for scales of 0.2 Mpc ⁻¹ ≤ k ₁ ≤ 0.9 Mpc ⁻¹ , with 100 hr of SKA-Mid observations per pointing. However, the detectability of the cross-bispectrum for large scales ( k ₁ < 0.1 Mpc ⁻¹ ), which is accessible with the single-dish mode of the survey, is limited by cosmic variance. Additionally, the signal loss due to foreground removal further suppresses the detectability. Our analysis presents a first step toward an end-to-end analysis pipeline for the future 21 cm-galaxy cross-bispectrum observations.

The evolution of galaxies is profoundly influenced by the environment they reside in. Cosmic voids, the most underdense regions of the Universe, serve as pristine laboratories for studying galaxy evolution in the relative absence or weaker influence of the complex physical processes that dominate denser environments. We investigate the properties and merger histories of galaxies as a function of environment using the GAlaxy Evolution and Assembly (GAEA) mock-observation light cone, which replicates the Euclid Deep Survey as foreseen at the epoch of the first data release. The galaxy sample spans the redshift range 0.4<z<1.8, corresponding to the interval over which is accessible to Euclid slitless spectroscopy. We classified galaxies based on their void-centric distance and local density contrast. We compared their stellar masses, specific star formation rates, bulge-to-total stellar mass ratios, and halo masses across different environments, while controlling for the dependence on stellar mass of the other intrinsic properties. We further analysed the merger histories of these galaxies to study their assembly evolution. We find that galaxies located closer to void centres (d_ ̊m cc ̊m v ) are generally less massive, more actively star-forming, and more disc-dominated than galaxies in denser regions, with their distributions statistically distinct at more than 4 confidence according to Kolmogorov--Smirnov tests. Merger histories indicate that void galaxies do not experience fewer mergers but that these occur later, leading to stellar mass assembly delayed by approximately 0.5--1 Gyr relative to galaxies in high-density regions. This supports a scenario in which the environment regulates the timing and nature of mergers rather than their overall frequency, producing a slower evolutionary path in low-density regions. We conclude by discussing the extent to which these trends are shaped by environmental parametrisation methods and observational selection effects. Our analysis provides a framework for interpreting forthcoming Euclid data and demonstrates that Euclid will enable the identification of cosmic voids and characterisations of their galaxy populations with an unprecedented statistical power and redshift coverage, providing a quantitative forecast of Euclid's capability to probe environmental effects on galaxy evolution.

Context. H I intensity mapping is a promising technique to probe large-scale structure, traditionally analyzed via two-point statistics such as the angular power spectrum. This latter technique has proven very powerful but may miss key non-Gaussian information present in the signal. Aims. We extend the starlet ℓ ₁ -norm, a multi-scale higher-order statistic previously applied to weak lensing maps, to the brightness temperature fluctuations of the H I density field. The H I signal is highly non-Gaussian at late times ( z < 1) due to nonlinear structure growth, motivating the use of advanced summary statistics. Methods. We simulated full-sky H I lognormal brightness temperature maps using CAMB and GLASS, generating 10 000 realizations with associated cosmological parameters. We extracted both the starlet ℓ ₁ -norm and angular power spectrum from these maps. Using the JaxILI framework, we performed neural density estimation for implicit likelihood inference. The analysis considered simulated maps incorporating realistic noise and telescope beam, capturing the impact of observational effects on parameter inference. In this work, we focus on the redshift range 0.4 < z < 0.45, chosen to match the interval already targeted by existing MeerKLASS observations. We also assess the sensitivity of these statistics to observational systematics, highlighting their potential for identifying and mitigating contaminants in H I intensity maps. Results. The starlet ℓ ₁ -norm significantly outperforms the angular power spectrum in constraining cosmological parameters, achieving almost a 3× improvement in the figure of merit relative to the angular power spectrum by capturing non-Gaussian features missed by two-point statistics. Moreover, our results suggest that the starlet ℓ ₁ -norm is robust to several of the systematic effects included in our simulations. Conclusions. Our findings highlight the potential of multi-scale higher-order statistics such as the starlet ℓ ₁ -norm to enhance cosmological inference from future H I intensity mapping surveys.

Recent experiments in cosmology, particularly those aimed at detecting the faint, redshifted, global 21[ ]cm hydrogen line (depth [ ], [ ]), have imposed stringent new requirements on radiometer calibration. In this work, we present a framework for circuit modeling and parameter inference to strengthen these calibration pipelines. This new approach enables in situ characterization of otherwise immeasurable systematics using physically motivated models. A combination of frequentist and Bayesian techniques are employed in a pipeline that supports iterative modeling, robust parameter estimation, and detailed uncertainty quantification. The framework is applied to the REACH telescope, where the precise correction of variations in the radio signal paths arising from component aging or environmental effects is critical. Circuit models of REACH’s calibration sources are developed, with the goal of predicting source temperature corrections that are conventionally obtained from laboratory measurements. By fitting the models to measured data using a convolutional cost function, a strong agreement with RMS residuals no worse than [ ] is obtained. However, Bayesian inference reveals that the resulting temperature corrections can have uncertainties on the order of 1 to [ ], caused by reflection coefficient degeneracies, measurement noise, and errors in the models. To combat this, posteriors obtained from laboratory measurements are employed as updated priors, reducing correction uncertainties down to [ ]. Ultimately, the framework provides a means of dynamically accounting for drift in system non-idealities over time, addressing the increasing precision demands of global 21 cm radio astronomy.

The ongoing Euclid mission is aimed at measuring spectroscopic redshifts for approximately two million galaxies using the H α line emission detected in near-infrared slitless spectroscopic data from the Euclid Deep Fields, leveraging both the red and blue grisms. These measurements will reach a flux limit of 5 × 10 ⁻¹⁷ erg cm ⁻² s ⁻¹ in the redshift range 0.4 < z < 1.8, paving the way to numerous scientific investigations involving galaxy evolution, extending well beyond the mission’s core objectives. The achieved H α luminosity depth will lead to a sufficiently high sampling, enabling the reconstruction of the large-scale galaxy environment. Here, we assess the quality of the reconstruction of the galaxy cosmic web environment with the expected spectroscopic dataset in Euclid Deep Fields. The analysis was carried out on the Flagship and GAEA galaxy mock catalogues. The quality of the reconstruction was first evaluated using simple geometrical and topological statistics measured on the cosmic web network; namely, the length of filaments, the area of walls, the volume of voids, and its connectivity and multiplicity. We then quantified how accurately gradients in galaxy properties can be recovered, with respect to the distance from filaments. As expected, the small-scale redshift-space distortions, such as Fingers of God (FoG) effects, have a strong impact on filament lengths and connectivity; however, they can be mitigated by compressing galaxy groups identified with an anisotropic group finder prior to a skeleton extraction. The cosmic web reconstruction is biased when relying solely on H α emitters. This limitation can be mitigated by applying stellar mass weighting during the cosmic web reconstruction. However, this approach introduces non-trivial biases that need to be accounted for when comparing to theoretical predictions. Redshift uncertainties pose the greatest challenge in recovering the expected dependence of galaxy properties, although the well-established stellar mass transverse gradients towards filaments can still be observed to a lesser extent.

Estimation des paramètres cosmologiques à partir des données du satellite Planck : de la construction d’une fonction de vraisemblance aux caractéristiques des neutrinos: Le fond diffus cosmologique (CMB), relique du Big-Bang chaud, porte les traces à la fois de la formation des structures des époques récentes et des premières phases énergétiques de l'Univers.Le satellite Planck, en service de 2009 à 2013, a fourni des mesures de haute qualité des anisotropies du CMB. Celles-ci sont utilisés dans cette thèse pour déterminer les paramètres du modèle cosmologique standard et autour du secteur des neutrinos.Ce travail décrit la construction d'un fonction de vraisemblance pour les hauts-l de Planck. Cela implique une stratégie de masquage pour l'émission thermique de la Galaxie ainsi que pour les sources ponctuelles. Les avant-plans résiduels sont traités directement au niveau du spectre de puissance en utilisant des templates physiques bases sur des études de Planck.Les méthodes statistiques nécessaires pour extraire les paramètres cosmologiques dans la comparaison entre les modèles et les données sont décrites, à la fois la méthode bayésienne de Monte-Carlo par chaînes de Markov et la technique fréquentiste du profil de la fonction de vraisemblance.Les résultats sur les paramètres cosmologiques sont présentés en utilisant les données de Planck seul et en combinaison avec les données à petites échelles des expériences de CMB basées au sol (ACT et SPT), avec les contraintes provenant des mesures des oscillations acoustiques des baryons (BAO) et des supernovae. Les contraintes sur l'échelle absolue de la masse des neutrinos et sur le nombre effectif de neutrinos sont également discutées.

Upcoming galaxy surveys with the SKA Observatory will detect neutral hydrogen (HI) across unprecedented volumes, and their scientific return will crucially depend on predictive models for HI observables. In this work, we present a framework to simulate the neutral hydrogen 21cm emission line in such large-scale HI galaxy surveys. This framework is developed as a modular layer that builds on semi-analytical models. In particular we use as bases the Galaxy Evolution and Assembly (GAEA) and L-Galaxies semi-analytical models, coupled to merger trees from the Millennium Simulation suite. We validate our framework against local Universe observations, demonstrating consistency with velocity functions, and generalised Tully-Fisher relations. Predictions based on GAEA and L-Galaxies exhibit mutual consistency despite the distinct underlying physical prescriptions. We construct mock galaxy catalogues that incorporate forward-modelled selection functions, inclination effects, and redshift broadening, reproducing the statistical distributions of HI-selected galaxies in the ALFALFA survey. Finally, we present redshift distribution forecasts for future SKA Observatory HI galaxy surveys. This framework offers a flexible tool for interpreting upcoming large-scale radio surveys and studying HI line observables as cosmological probes.