Yurii Kvasiuk

Yurii Kvasiuk

PhD student in Physics

KSZ Velocity Reconstruction with ACT and DESI‑LS using a Tomographic QML Power Spectrum Estimator

Overview

This project performs the kinetic Sunyaev‑Zel’dovich (kSZ) velocity reconstruction using real observational data from ACT DR6 and DESI‑LS DR9. We use the quadratic estimator to reconstruct velocity. A new tomographic Quadratic Maximum Likelihood (QML) estimator is developed, operating in redshift‑binned spherical coordinates (onion picture), to cross correlate the reconstructed velocity field with galaxy density. We find that widely assumed Battaglia-AGN profile of the electron density results in the \(A=0.39\) for the reconstructed amplitude of the velocity, possibly indicating the presence of enhanced feedback

Methodology

  • Data Sets
    • ACT (Atacama Cosmology Telescope) Data Release 6 (DR6) for CMB temperature maps.
    • DESI‑LS (Dark Energy Spectroscopic Instrument – Large Survey) Data Release 9 galaxy catalog.
  • Estimator / Reconstruction
    • A quadratic estimator (QE) is used to reconstruct the velocities (note: even though it is suboptimal, for the given experimental noises it’s enough. However, for the futuristic experiments, one could benefit from the field likelihood analysis)
    • We develop a new tomographic QML (Quadratic Maximum Likelihood) in spherical coordinates with bins in redshift. This allows to optimally estimate the cross‑powers between velocity field and galaxy density on the lightcone

Results

We report an 11.7 σ detection of the cross-correlation between the reconstructed kSZ velocity field (from ACT DR6) and the DESI-LS DR9 galaxy density field.
Central to this achievement is the introduction of a tomographic Quadratic Maximum Likelihood (QML) estimator, which a general optimal power spectrum estimator for an arbitrary \(N\) correlated Gaussian fields on a sphere.

A key physical result is the recovered amplitude parameter (A \approx 0.4) relative to standard halo-model predictions. This value well below unity indicates strong astrophysical feedback, such as gas heating or ejection from massive halos, which suppresses the expected kSZ signal relative to the Battaglia-AGN electron density profile.

Future Prospects

In the future, we plan to use the reconstructed velocity to obtain the cosmological parameter constraints, in particular the local non-Gaussianity \(f_{NL}\), and possibly others (\(m_\nu,\frac{dw}{da}\))

  • Paper: KSZ Velocity Reconstruction with ACT and DESI‑LS using a Tomographic QML Power Spectrum Estimator — arXiv: 2506.21684