Mesh-free free-form lensing I: Methodology and application to mass reconstruction

TL;DR

This paper introduces a mesh-free numerical framework based on radial basis functions for gravitational lensing data analysis, enabling accurate mass reconstruction across different regimes and promising application to real observational data.

Contribution

The authors develop a novel mesh-free RBF-based interpolation and differentiation method for gravitational lensing, improving mass reconstruction accuracy without relying on regular meshes.

Findings

Achieves sub-percent interpolation and differentiation accuracy.

Combining strong and weak lensing constraints yields full-field accurate mass reconstructions.

Method performs well on simulated data, promising for real observational applications.

Abstract

Many applications and algorithms in the field of gravitational lensing make use of meshes with a finite number of nodes to analyze and manipulate data. Specific examples in lensing are astronomical CCD images in general, the reconstruction of density distributions from lensing data, lens-source plane mapping or the characterization and interpolation of a point-spread-function. We present a numerical framework to interpolate and differentiate in the mesh-free domain, defined by nodes with coordinates that follow no regular pattern. The framework is based on radial basis functions (RBFs) to smoothly represent data around the nodes. We demonstrate the performance of Gaussian RBF-based, mesh-free interpolation and differentiation, which reaches the sub-percent level in both cases. We use our newly developed framework to translate ideas of free-form mass reconstruction from lensing onto the mesh-free domain. By reconstructing a simulated mock lens we find that strong lensing only reconstructions achieve < 10% accuracy in the areas where these constraints are available but provide poorer results when departing from these regions. Weak-lensing only reconstructions give < 10% percent accuracy outside the strong lensing regime, but cannot resolve the inner core structure of the lens. Once both regimes are combined, accurate reconstructions can be achieved over the full field of view. The reconstruction of a simulated lens, using constraints that mimics real observations, yields accurate results in terms of surface-mass density, NFW parameter, Einstein radius and magnification map recovery, encouraging the application of this method to real data.