Controlling Data Systematics in Euclid Spectroscopy with Simulations and Observations
Passalacqua, Francesca; Anselmi, Stefano; dusini, stefano; Renzi, Alessandro; SIRIGNANO, Chiara; TROJA, ANTONINO
- Publisher
- Zenodo
- Language
- en
Abstract
Parallel talk presented at the XXI International Workshop on Neutrino Telescopes - Padova 29 September - 3 October 2025 (https://agenda.infn.it/event/44606/) Abstract: Euclid is a mission of the European Space Agency (ESA), designed to investigate the content and evolution of the Universe. Launched in July 2023, the satellite will collect data for at least six years, covering one-third of the sky. Euclid holds the promise to provide crucial new constraints on relevant cosmological parameters, including the sum of the neutrino masses. Accurate cosmological inference from Euclid data requires careful treatment of observational systematics, which could bias the measured galaxy distribution and distort parameter estimation. Galaxy clustering analyses, in particular, depend on a detailed understanding of the purity and completeness of the cosmological sample. The idea is to assess these properties using two complementary approaches: simulations and deep observations of reference fields. While simulations offer a controlled environment with known inputs, they must be computationally efficient and account for instrumental effects that are not fully understood. Indeed, certain detector non-idealities, such as persistence and snowballs, remain difficult to model accurately, limiting the realism of such simulations. However, deeper observations are confined to small sky regions and present complex selection functions, limiting their representativeness of the data sample. We first present a fully simulated pipeline to assess the systematic effects, such as spectral overlap, which is one of the main challenges of Euclid's slitless spectroscopic technique. We also present a complementary approach in which we simulate only the spectra; those are then injected into real Euclid images and then processed through the full data pipeline. This method naturally inherits instrumental systematics from real data, offering a more realistic characterisation of purity and completeness. Though computationally intensive, it enables the assessment of systematics over larger sky areas and provides a valuable complementary strategy for robust cosmological analyses and neutrino mass investigations.
Full text
F. Passalacqua, S. Anselmi, S. Dusini, A. Renzi, C. Sirignano, A. Troja Controlling Data Systematics in Euclid Spectroscopy with Simulations and Observations 1/24
1. Galaxy Clustering 2. Slitless spectroscopy 3. Assessment of observational systematics in Euclid Overview 2/24
Observations Redshift measurement Sample selection Compute some statistics Cosmological inference Cosmology from Galaxy Clustering 3/24
Observations Redshift measurement Sample selection Compute some statistics Cosmological inference Cosmology from Galaxy Clustering 3/24 Euclid Collaboration: Le Brun et al., arXiv
Observations Redshift measurement Sample selection Compute some statistics Cosmological inference Cosmology from Galaxy Clustering 3/24 Euclid Collaboration: Le Brun et al., arXiv
Observations Redshift measurement Sample selection Compute some statistics Cosmological inference Cosmology from Galaxy Clustering 3/24 Euclid Collaboration: Le Brun et al., arXiv
Observations Redshift measurement Sample selection Compute some statistics Cosmological inference Cosmology from Galaxy Clustering 3/24 Euclid Collaboration: Le Brun et al., arXiv
Computing Clustering Statistics 4/24
•Ground-based •Optical wavelengths (360 to 980 nm) •Fiber •Wavelength resolution R = λ/Δλ= 2000 - 5000 •Photometric sample selection →a priori •Space-based •NIR wavelengths (1200 to 1850 nm) •Slitless spectroscopy •Wavelength resolution = R = λ/Δλ> 380 •“Spectroscopic” sample selection →a posteriori DESI Euclid Stage IV Surveys –a comparison → Different systematics 5/24
Drawbacks 1. High background level 2. Self-contamination: mixing of the spatial and spectral information →the effective LSF depends on the shape of the source along the dispersion direction 12/24
Drawbacks 1. High background level 2. Self-contamination 3. Cross-contamination: overlap of light from different sources (mainly 0th and 1st order in Euclid) 13/24
Assessment of observational systematics in Euclid 14/24
Understanding systematics: 1. Simulations →study simple effects 2. Use deeper observations →Euclid Deep Survey 3. … Evaluation of systematics 15/24
Spectroscopic Simulations Measurement of the redshift success rate as a function of line flux and size of the galaxies . Euclid Collaboration: Passalacqua et al., in prep 16/24
Understanding Systematics with Observations Assessment of Purity and Completeness Euclid deep fields: –Required depth will be reached only at the end of the survey –Assessment of P&C of deep data 17/24
Understanding Systematics with Observations Assessment of Purity and Completeness Euclid deep fields: –Required depth will be reached only at the end of the survey –Assessment of P&C of deep data Modelling systematics 1. Known: –Light background – Thermal, readout noise, … 2. Known unknowns: –Persistence –Residuals from crosscontamination (0th, 1st, 2nd orders) –Missing or fake objects 3. Unknown unknowns… 18/24
Spectroscopic Source Injection Idea: study unknown systematics by combining pros of simulations and observed data: •Simulations →known input •Real observations →contain systematics →Inherit systematic from observed data and learn from the data →Reprocess the spectra and measure the redshift with the Euclid pipelines 19/24
1. Retrieve pre-processed observation Spectroscopic Source Injection 2. Simulate and inject fake spectra * Gelsa code (B. Granett) 20/24
Correct emission line identification when injected Examples 21/24
