Multiple component decomposition from millimeter single-channel data

TL;DR

This paper introduces a blind source separation method for millimeter single-channel data, enabling the decomposition of complex signals into physical components, demonstrated on simulated and real survey data.

Contribution

The authors develop a novel decomposition technique that generates artificial redundancy to separate multiple physical components from single-wavelength millimeter survey data.

Findings

Successfully decomposed survey data into four physical components

Reduced flux bias and improved signal-to-noise ratio

Effectively separated atmospheric, systematic, and astrophysical signals

Abstract

We present an implementation of a blind source separation algorithm to remove foregrounds off millimeter surveys made by single-channel instruments. In order to make possible such a decomposition over single-wavelength data: we generate levels of artificial redundancy, then perform a blind decomposition, calibrate the resulting maps, and lastly measure physical information. We simulate the reduction pipeline using mock data: atmospheric fluctuations, extended astrophysical foregrounds, and point-like sources, but we apply the same methodology to the AzTEC/ASTE survey of the Great Observatories Origins Deep Survey-South (GOODS-S). In both applications, our technique robustly decomposes redundant maps into their underlying components, reducing flux bias, improving signal-to-noise, and minimizing information loss. In particular, the GOODS-S survey is decomposed into four independent physical components, one of them is the already known map of point sources, two are atmospheric and systematic foregrounds, and the fourth component is an extended emission that can be interpreted as the confusion background of faint sources.