Beyond Point Photo- z Estimates: Normalizing Flows for Tomographic N(z) Reconstruction with DESI DR1
astro-ph.CO
Submitted: 2026-10-29
Updated: 2026-10-29
Code: https://github.com/chenzhuo0623-pixel/Normalizing-flowphotometric-redshift-PDFs-with-DESI-DR1
Project page: https://zillow.github.io/quantile-forest
License: http://creativecommons.org/licenses/by/4.0/
The gist: Accurate calibration of tomographic redshift distributions, N(z) is essential for cosmological analysis of weak lensing data.
Terminology
Abstract
Accurate calibration of tomographic redshift distributions, N(z) is essential for cosmological analysis of weak lensing data. We investigate conditional normalizing flows (NFs) and compare their performance with random forest (RF) and quantile random forest (QRF) models. We consider DESI galaxy and quasar targets from DESI DR1 spectroscopic data and with the photometry from Legacy Surveys DR10. We train on 1.1 million sources and test on 2.0 million sources in a non-overlapping sky region (0<z spec<4.5). For point estimation, the NF achieves the lowest scatter and outlier fraction at 0<z spec<1.6 and outperforms both baselines across all three metrics at z spec>1.6. By introducing distribution-level metrics like Wasserstein distance and JS divergence, we find that stacking NF posterior PDFs provides more accurate tomographic N(z) reconstruction than QRF, particularly at high redshifts. For z median>2.5, the NF achieves a smaller absolute mean-redshift bias (0.060 plus or minus0.031 versus 0.318) and higher probability overlap (0.905 plus or minus0.008 versus 0.787) than QRF. We apply redshift-dependent post-hoc PDF calibration derived from a held-out validation sample. After calibration, the absolute mean-redshift biases in four bins spanning 0.2<z median<1.0 range from 0.0016 to 0.0019, with central values below referring benchmark 0.002(1+z). In the highest-redshift bin, calibration further reduces the bias to 0.016 plus or minus0.011. Our results demonstrate the potential of NF PDFs for tomographic N(z) reconstruction while highlighting challenges in transferring models across galaxy populations.
Sources
- Target Selection and Validation of DESI Quasars
- Re-calibrating Photometric Redshift Probability Distributions Using Feature-space Regression
- NICE: Non-linear Independent Components Estimation
- Euclid Definition Study Report
- Clustering-based redshift estimation: method and application to data
- Zephyr : Stitching Heterogeneous Training Data with Normalizing Flows for Photometric Redshift Inference
- Photometric Redshift PDFs via Neural Network Classification for DESI Legacy Imaging Surveys and Pan-STARRS
Related papers
- 21cmEMUv3: a hybrid diffusion-LSTM emulator of 21cmFAST summary observables
- Connecting field-level and n-point function measurements of galaxy bias
- The Impact of Cosmic Variance and Satellites on JWST Clustering Measurements at Redshift around 6
- Universality of Primordial Anisotropies in Gravitational Wave Background
- KiDS-1000 cosmic shear reanalysis using MetaCalibration
- 2Flow2SN: Dusty Supernova Cosmology with Photometric Redshifts and Contamination using Neural Simulation-Based Inference