Exploration of 3D wavelet scattering transform coefficients for line-intensity mapping measurements

TL;DR

This paper investigates the use of 3D wavelet scattering transform coefficients for analyzing line-intensity mapping data, demonstrating their potential to extract non-Gaussian features and improve cosmological parameter constraints beyond traditional methods.

Contribution

It introduces the application of 3D solid harmonic WST to line-intensity mapping, showing its interpretability, potential for enhanced constraints, and the importance of hyperparameter tuning and covariance estimation.

Findings

WST coefficients are interpretable in noiseless simulations.

WST can detect the cosmological signal at $z\,\sim3$ even in early phases.

Reduced WST sets can outperform power spectrum constraints.

Abstract

The wavelet scattering transform (WST) has recently gained attention in the context of large-scale structure studies, being a possible generator of summary statistics encapsulating non-Gaussianities beyond the reach of the conventional power spectrum. This work examines the three-dimensional solid harmonic WST in the context of a three-dimensional line-intensity mapping measurement to be undertaken by current and proposed phases of the CO Mapping Array Project (COMAP). The WST coefficients demonstrate interpretable behaviour in the context of noiseless CO line-intensity simulations. The contribution of the cosmological $z\sim3$ signal to these coefficients is also detectable in principle even in the Pathfinder phase of COMAP. Using the peak-patch method to generate large numbers of simulations and incorporating observational noise, we numerically estimate covariance matrices and show that careful choices of WST hyperparameters and rescaled or reduced coefficient sets are both necessary to keep covariances well-conditioned. Fisher forecasts show that even a reduced `shapeless' set of $\ell$-averaged WST coefficients show constraining power that can exceed that of the power spectrum alone even with similar detection significance. The full WST could improve parameter constraints even over the combination of the power spectrum and the voxel intensity distribution, showing that it uniquely encapsulates shape information about the line-intensity field. However, practical applications urgently require further understanding of the WST in key contexts like covariances and cross-correlations.