Francesco Pistis

Context. Current and future large spectroscopic surveys are significantly enhancing the volume and resolution of quasar spectra that are observed, motivating the creation of efficient and precise automated techniques to detect absorption features. Aims. This study focuses on the detection of metal absorption features using a novel U-Net model on WEAVE-like mock spectra in the quasar rest-frame wavelength interval 1230 Å ≤ λ _RF ≤ 3095 Å. We test the network performance for absorption detection both on ideal data and after simulating the continuum fitting step as applied on real data. The ability to recover the correct line centers is also studied. Methods. The performance of these architectures is evaluated by the completeness, purity, and F1 score reached in bins of signal-to-noise (S/N _line ) for the absorption lines and with the absolute fractional flux error for the continuum. Results. The U-Net model reaches scores of ≈90% for all metrics (completeness, purity, and F1 score) at S/N _line ≈ 4. All false positive detections with S/N _line ≥ 5 fall in the tails of the broad Ly α absorbers distribution of damped Ly α systems. The combination of continuum fitting and line detections has negligible effects on the detection performance at S/N _line ≥ 4. Conclusions. Our proposed U-Net architecture offers a competitive tool for the analysis of absorption lines in current and upcoming large spectroscopic surveys and is well suited to the identification of any absorption line feature.

We present the first detailed forecasts for the detectability of patchy hydrogen reionization in the one-dimensional Lyα forest power spectrum to be measured by the WEAVE-QSO survey. Using the Sherwood-relics reionization simulations and a WEAVE-QSO survey configuration, we generate mock spectra in four redshift bins, z = 4.0, 4.2, 4.4, and 4.6, in which relic ionization and temperature fluctuations from patchy hydrogen reionization enhance the Lyα forest power spectrum on large scales (i.e., at wavenumber $k\sim 10^{-3}\rm \, s\, km^{-1}$). Our Lyα forest pipeline forecasts the power spectrum covariance by considering sample size, spectral resolution, noise subtraction, continuum placement, metal contamination, and damping wings from high-column density absorbers. Applying our covariance forecast within a Bayesian parameter inference framework, we find that the signature of patchy hydrogen reionization should be detectable at a significance of ≃ 4.5σ. The forthcoming WEAVE-QSO 1D power spectrum measurements should therefore be able to directly detect and characterize the large-scale relic imprint of patchy hydrogen reionization in the Lyα forest power spectrum at z ≥ 4.

We present a deep learning-based framework for automated quasar spectral analysis, addressing two key challenges: continuum estimation and absorption feature detection. Using autoencoders, convolutional neural networks (CNNs), and U-Net architectures, we demonstrate improved performance in reconstructing the quasar continuum and identifying absorption lines. These models are trained to learn complex, non-linear correlations across wide spectral regions, enabling robust predictions even in the presence of noise or missing data. Our methods offer greater flexibility and accuracy, particularly in low signal-to-noise regimes. The U-Net architecture, in particular, shows strong potential for precise absorption feature detection, leveraging its encoder-decoder structure with skip connections to capture both local and global spectral patterns. This enables the identification of absorption signatures associated with metal-enriched gas in the circumgalactic medium. Our results highlight the scalability of these techniques for application to large-scale surveys such as DESI and WEAVE, making progress for new insights into the distribution and dynamics of baryons in the Universe.

Context. The new generation of optical spectrographs, including WEAVE, 4MOST, DESI, and WST, offers huge multiplexing capabilities and excellent spectral resolution. This is an unprecedented opportunity to statistically unveil the details of the star formation histories (SFHs) of galaxies. However, these observations are not easily comparable with the predictions of cosmological simulations. Aims. Our goal is to build a reference framework for comparing spectroscopic observations with cosmological simulations and to test the currently available tools for deriving the stellar population properties of mock galaxies as well as their SFHs. We focus on the observational strategy of the Stellar Population at intermediate redshift Survey (StePS) carried out with the WEAVE instrument. Methods. We created mock datasets of ∼750 galaxies at redshifts of z = 0.3, 0.5, and 0.7 from the TNG50 cosmological simulation. We performed radiative transfer calculations using SKIRT and analyzed the spectra with the pPXF algorithm, treating them as if they were real observations. Results. This work presents the methodology used to generate the mock datasets, providing an initial exploration of stellar population properties (i.e. mass-weighted ages and metallicities) and SFHs of a test sample of three galaxies at z = 0.7 and their descendants at z = 0.5 and 0.3. We show that there is very good agreement between mock WEAVE-like spectra compared to the intrinsic values in TNG50 (average difference of 0.2 ± 0.3 Gyr). We also report that an overall agreement is seen when retrieving the SFHs of galaxies, especially if they form the bulk of their stars on short timescales and at early epochs. While we did identify a tendency to overestimate the weight of old stellar populations in galaxies with complex SFHs, we were able to properly recover the timescales on which galaxies build up 90% of their mass, with almost no difference in the measured and intrinsic cumulative SFHs over the last 4 Gyr. Conclusions. We have released the datasets concurrently with this publication of this paper, which consist of multi-wavelength imaging and spectroscopic data of ∼750 galaxies at redshift z = 0.3, 0.5, and 0.7. This work provides a fundamental bench-test for forthcoming WEAVE observations, providing the community with realistic mock spectra of galaxies that can be used to test currently available tools for deriving first-order stellar populations parameters (i.e. ages and metallicities) as well as more complex diagnostics, such as mass and SFHs.

Context. Ongoing and upcoming large spectroscopic surveys are drastically increasing the number of observed quasar spectra, making the development of fast and accurate automated methods to estimate spectral continua necessary. Aims. This study evaluates the performance of three neural networks (NNs) – an autoencoder, a convolutional NN (CNN), and a U-Net – in predicting quasar continua within the rest frame wavelength range of 1020 Å to 2000 Å. The ability to generalize and predict galaxy continua within the range of 3500 Å to 5500 Å is also tested. Methods. We evaluated the performance of these architectures using the absolute fractional flux error (AFFE) on a library of mock quasar spectra for the WEAVE survey and on real data from the early data release observations of the Dark Energy Spectroscopic Instrument (DESI) and the VIMOS Public Extragalactic Redshift Survey (VIPERS). Results. The autoencoder outperforms U-Net, achieving a median AFFE of 0.009 for quasars. The best model also effectively recovers the Lyα optical depth evolution in the DESI quasar spectra. With minimal optimization, the same architectures can be generalized to the galaxy case, with the autoencoder reaching a median AFFE of 0.014 and reproducing the D4000n break in DESI and VIPERS galaxies.

Aims. Large future sky surveys, such as the Legacy Survey of Space and Time (LSST), will provide optical photometry for billions of objects. Reliable estimation of the physical properties of galaxies requires information about dust attenuation, which is usually derived from ultraviolet (UV) and infrared (IR) data. This paper aims to construct a proxy for the far-UV (FUV) attenuation (A_FUVp) from the optical data alone, enabling the rapid estimation of the star formation rate (SFR) for galaxies that lack UV or IR data. This will accelerate and improve the estimation of key physical properties of billions of LSST–like observed galaxies (observed in the optical bands only). Methods. To mimic LSST observations, we used the deep panchromatic optical coverage of the Sloan Digital Sky Survey (SDSS) Photometric Catalogue, Data Release 12, complemented by the estimated physical properties for the SDSS galaxies from the GALEX-SDSS-WISE Legacy Catalog (GSWLC) and inclination information obtained from the SDSS Data Release 7. We restricted our sample to the 0.025–0.1 spectroscopic redshift range and investigated relations among surface brightness, colours, and dust attenuation in the FUV range for star-forming galaxies obtained from the spectral energy distribution (SED). Results. Dust attenuation is best correlated with colour measured between u and r bands (u − r) and the surface brightness in the u band (μ_u). We provide a dust attenuation proxy for galaxies on the star-forming main sequence. This relation can be used for the LSST or any other type of broadband optical survey. The mean ratio between the catalogue values of SFRs and those estimated using optical-only SDSS data with the A_FUVp prior calculated as ΔSFR = log(SFR_this work/SFR_GSWLC) is found to be less than 0.1 dex, while runs without priors result in an SFR overestimation larger than 0.3 dex. The presence or absence of the A_FUVp has a negligible influence on the stellar mass (M_star) estimation (with ΔM_star in the range from 0 to −0.15 dex). Conclusions. We note that A_FUVp is reliable for low-redshift main sequence galaxies. Forthcoming deep optical observations of the LSST Deep Drilling Fields, which also have multi-wavelength data, will enable one to calibrate the obtained relation for higher redshift galaxies and, possibly, extend the study towards other types of galaxies, such as early-type galaxies off the main sequence.

Aims. We aim to investigate the influences on the evolution of the fundamental metallicity relation of different selection criteria. Methods. We used 5 487 star-forming galaxies at a median redshift of z ≈ 0.63 extracted from the VIMOS Public Extragalactic Redshift Survey (VIPERS) and 143 774 comparison galaxies in the local Universe from the GALEX-SDSS-WISE Legacy Catalog. We employed two families of methods: parametric and nonparametric. In the parametric approaches, we compared the fundamental metallicity relation projections plagued by observational biases on differently constructed control samples at various redshifts. Then, we compare the metallicity difference between different redshifts in stellar mass-star formation rate bins. In the nonparametric approach, we related the metallicity and the normalized specific star formation rate (sSFR). To compare galaxies with the same physical properties, we normalized the sSFR of our samples according to the median value at median redshift z ≈ 0.09. Then, the galaxies with the same distance from the star-forming main sequence at their respective redshifts were compared when the sSFR is normalized according to the expected values from their respective star-forming main sequences. Results. The methodologies implemented to construct fair, complete samples for studying the mass-metallicity relation and the fundamental metallicity relation produced consistent results showing a small but still statistically significant evolution of both relations up to z ≈ 0.63. In particular, we observed a systematic trend where the median metallicity of the sample at z = 0.63 is lower than that of the local sample at the same stellar mass and star formation rate. The average difference in the metallicity of the low and intermediate redshifts is approximately 1.8 times the metallicity standard deviation of the median of the intermediate redshift sample in stellar mass-star formation rate bins. We confirmed this result using the Kolmogorov-Smirnov test. When we applied the stellar mass-completeness criterion to catalogs, the metallicity difference in redshifts decreased to approximately 0.96 times the metallicity standard deviation of the median, and thus it was not statistically significant. The limited area of the FMR surface explored once the stellar mass-completeness criterion is applied might dominate this difference reduction, leaving out the area where the difference between the two samples is the highest. A careful reading of the results and their underlying selection criteria is crucial in studies of the mass-metallicity and fundamental metallicity relations. Conclusions. When studying the mass-metallicity and fundamental metallicity relations, we recommend using the nonparametric approach, which provides similar results compared to parametric prescriptions, is easier to use, and is clear to interpret. The nonparametric methodology provides a convenient way to compare physical properties, with a smaller impact on observational selection biases.

Aims. This work aims to determine how the galaxy main sequence (MS) changes using seven different commonly used methods to select the star-forming galaxies within VIPERS data over 0.5 ≤ z < 1.2. The form and redshift evolution of the MS was then compared between selection methods. Methods. The star-forming galaxies were selected using widely known methods: a specific star-formation rate (sSFR); Baldwin, Phillips, and Terlevich (BPT) diagram; a 4000 Å spectral break (D4000) cut; and four colour-colour cuts (near-ultra-violet – V verses r − J (NUVrJ), near-ultra-violet – V verses r − K (NUVrK), u − r, and U − V verses V − J (UVJ)). The main sequences were then fitted for each of the seven selection methods using a Markov chain Monte Carlo forward modelling routine, fitting both a linear main sequence and a MS with a high-mass turnover to the star-forming galaxies. This was done in four redshift bins of 0.50 ≤ z < 0.62, 0.62 ≤ z < 0.72, 0.72 ≤ z < 0.85, and 0.85 ≤ z < 1.20. Results. The slopes of all star-forming samples were found to either remain constant or increase with redshift, and the scatters were approximately constant. There is no clear redshift dependency of the presence of a high-mass turnover for the majority of samples, with the NUVrJ and NUVrK being the only samples with turnovers only at low redshift. No samples have turnovers at all redshifts. Star-forming galaxies selected with sSFR and u − r are the only samples to have no high-mass turnover in all redshift bins. The normalisation of the MS increases with redshift, as expected. The scatter around the MS is lower than the ≈0.3 dex typically seen in MS studies for all seven samples. Conclusions. The lack (or presence) of a high-mass turnover is at least partially a result of the method used to select star-forming galaxies. However, whether a turnover should be present or not is unclear.

Context. The star formation rate (SFR) is a key ingredient for studying the formation and evolution of galaxies. Being able to obtain accurate estimations of the SFR, for a wide range of redshifts, is crucial for building and studying galaxy evolution paths over cosmic time. Aims. Based on a statistical sample of galaxies, the aim of this paper is to constrain a set of SFR calibrators that are able to work in a large redshift range, from z = 0 to z = 0.9. Those calibrators will help to homogenize SFR estimations of star-forming galaxies and to remove any possible biases from the study of galaxy evolution. Methods. Using the VIMOS Public Extragalactic Redshift Survey (VIPERS), we estimated a set of SFR based on photometric and spectroscopic data. We used, as estimators, photometric bands from ultraviolet (UV) to mid-infrared (mid-IR), and the spectral lines Hβ, [O II]λ3727, and [O III]λ5007. Assuming a reference SFR obtained from the spectral energy distribution reconstructed with Code Investigating GALaxy Emission (CIGALE), we estimated the reliability of each band as an SFR tracer. We used the GALEX-SDSS-WISE Legacy Catalog (GSWLC, z < 0.3) to trace the dependence of these SFR calibrators with redshift. Results. The far and near UV (FUV and NUV, respectively), u-band and 24 μm bands, as well as L_TIR, are found to be good SFR tracers up to z ∼ 0.9 with a strong dependence on the attenuation prescription used for the bluest bands (scatter of SFR of 0.26, 0.14, 0.15, 0.23, and 0.24 dex for VIPERS, and 0.25, 0.24, 0.09, 0.12, and 0.12 dex for GSWLC). The 8 μm band provides only a rough estimate of the SFR as it depends on metallicity and polycyclic aromatic hydrocarbon properties (scatter of 0.23 dex for VIPERS). We estimated the scatter of rest-frame luminosity estimations from CIGALE to be 0.26, 0.14, 0.12, 0.15, and 0.20 dex for FUV, NUV, ugriz, K_s, and 8–24 μm-L_TIR. At intermediate redshift, the Hβ line is a reliable SFR tracer (scatter of 0.19 dex) and the [O II]λ3727 line gives an equally good estimation when the metallicity from the R₂₃ parameter is taken into account (0.17 for VIPERS and 0.20 dex for GSWLC). A calibration based on [O III] retrieves the SFR only when additional information such as the metallicity or the ionization parameter of galaxies are used (0.26 for VIPERS and 0.20 dex for GSWLC), diminishing its usability as a direct SFR tracer. Based on rest-frame luminosities estimated with CIGALE, we propose our own set of calibrations from FUV, NUV, u-band, 8, 24 μm, L_TIR, Hβ, [O II], and [O III].