Euclid Q1: photometric redshifts and physical properties with NNPZ + IRAC in the Euclid Deep Fields
Authors/Creators
-
1.
CEA Paris-Saclay
- 2. INAF-Osservatorio Astronomico di Bologna
-
3.
Osservatorio di Astrofisica e Scienza dello Spazio
Description
This record contains photometric redshifts and physical properties (stellar mass, SFR, age, attenuation, metallicity and rest-frame absolute magnitudes) for 16.9 million sources in the three Euclid Deep Fields (EDF-North, EDF-Fornax, EDF-South) of the Euclid Quick Data Release (Q1). They were obtained by re-running the NNPZ code on the Q1 MER photometry with the addition of IRAC 3.6 and 4.5 μm data, a new reference sample and physically motivated priors, as described in Euclid Collaboration: Enia et al. (2026), A&A, 711, A14.
Files:
– Euclid_Q1_EDFN.fits, Euclid_Q1_EDFS.fits, Euclid_Q1_EDFF.fits: full results for each field (weighted median, 16th/84th percentiles, mean and nearest-neighbour values);
– Euclid_Q1_Redux.fits: reduced catalogue of all fields, with coordinates, H_E magnitude, spectroscopic redshift when available and median values;
– Euclid_Q1_at_z_spec.fits: physical properties at fixed spectroscopic redshift for 49 381 sources with reliable external spectroscopic redshifts;
– README.md: full description of the method, columns, units and known caveats.
Please read the README before using the data. If you use them, please cite Euclid Collaboration: Enia et al. (2026) and this record.
Methods (English)
NNPZ was run on all Q1 MER sources with SPURIOUS_FLAG = 0, DET_QUALITY_FLAG < 4, MUMAX_MINUS_MAG > −2.6 and I_E < 25.5 or H_E < 24.5, using the Euclid, ground-based (UNIONS/DES) and IRAC photometry. The reference sample contains 1 490 150 Bruzual & Charlot (2003) models with delayed-τ star-formation histories. For each source the 80 nearest neighbours are weighted by their χ² distance and by a volume prior and an age prior. Physical properties at fixed spectroscopic redshift were obtained by splitting both the targets and the reference sample in redshift slices and keeping only the neighbours within a narrow window around z_spec. Photo-z uncertainties are underestimated by a factor of ~1.5; see the README and the paper for validation and caveats.