AMICO-WL: an optimal filtering algorithm for galaxy cluster detections with weak lensing

TL;DR

AMICO-WL is an advanced optimal filtering algorithm designed for efficient and accurate detection of galaxy clusters through weak lensing signals, tailored for upcoming large-scale surveys like Euclid and LSST-Rubin.

Contribution

This work introduces AMICO-WL, a novel extension of the AMICO algorithm, optimized for weak lensing data with parallel processing and improved detection purity and completeness.

Findings

Doubles completeness from 6.5% to 13.0% with foreground removal.

Reduces spurious detections significantly.

Effective on Euclid-like mock galaxy catalogues.

Abstract

The detection of galaxy clusters, the most massive bounded structures in the universe, is crucial for cosmological analysis. Weak lensing signals allow us to track the distribution of all (dark and baryonic) matter regardless of its observable electromagnetic properties. Upcoming wide-field surveys like Euclid and LSST-Rubin will provide enhanced shape measurements of billions of background galaxies, presenting an unparalleled opportunity to detect galaxy clusters on a vast cosmic scale. The immense data volume generated by these surveys will require efficient and accurate analysis techniques. In this work, we introduce AMICO-WL, an extension of the optimal filtering algorithm implemented in AMICO, a well-tested code developed for optical cluster detection. AMICO-WL implements a specific linear optimal matched filter for weak lensing data in the AMICO infrastructure, using parallelisation and adding an efficient signal-to-noise ratio thresholding approach to set a desired sample purity and a cleaning procedure to deal with blended detections. The algorithm has been tested on a 25 deg$^2$ field of Euclid-like mock galaxy catalogue with the simulated shear signal produced using DUSTGRAIN-$pathfinder$ past-light-cones. We implemented a foreground removal procedure based on different cuts of low redshift galaxies from the input catalogue. To evaluate the performance of the method, we used an efficient matching procedure based on the 'blinking' of the simulation's individual redshift lensing planes. Cross-matching the AMICO-WL detections with the dark matter halo sample in the simulation having $M_{200} > 5 \times 10^{13}$ $M_{\odot}/h$ and considering a purity level of $\sim70\%$, the application of the foreground removal doubles the completeness from $6.5\%$ to $13.0\%$ and at the same time produces a significant decrease of spurious detections.