relensing: Reconstructing the mass profile of galaxy clusters from gravitational lensing

TL;DR

Relensing is a Python package for modeling galaxy clusters via gravitational lensing, improving reconstruction accuracy and stability through smoothing techniques, and providing a versatile tool for the scientific community.

Contribution

It introduces a free-form, adaptive grid approach with smoothing to enhance galaxy cluster mass profile reconstructions from lensing data.

Findings

Smoothing improves shape and size recovery of mass profiles.

Achieves rms of ~0.16-0.17 arcsec on simulated clusters.

Outperforms similar existing reconstruction methods.

Abstract

In this work we present relensing, a package written in python whose goal is to model galaxy clusters from gravitational lensing. With relensing we extend the amount of software available, which provides the scientific community with a wide range of models that help to compare and therefore validate the physical results that rely on them. We implement a free-form approach which computes the gravitational deflection potential on an adaptive irregular grid, from which one can characterize the cluster and its properties as a gravitational lens. Here, we use two alternative penalty functions to constrain strong lensing. We apply relensing to two toy models, in order to explore under which conditions one can get a better performance in the reconstruction. We find that by applying a smoothing to the deflection potential, we are able to increase the capability of this approach to recover the shape and size of the mass profile of galaxy clusters, as well as its magnification map. This translates into a better estimation of the critical and caustic curves. The power that the smoothing provides is also tested on the simulated clusters Ares and Hera, for which we get an rms on the lens plane of ~0.17 arcsec and ~0.16 arcsec, respectively. Our results represent an improvement with respect to reconstructions that were carried out with methods of the same nature as relensing. At the same time, the smoothing also increases the stability of our implementation, and decreases the computation time. In its current state, relensing is available upon request.