A Forward, Analytic, Differentiable, Geometric (But Inflexible) Lens Model

TL;DR

This paper introduces an analytic, differentiable lens model for gravitational lensing that offers significant computational speed advantages and visual verification capabilities, despite some geometric limitations.

Contribution

The authors present a new forward, analytic, differentiable lens model that speeds up computations and allows quick visual checks, addressing limitations of traditional models.

Findings

Speed-up factor exceeding 10,000 compared to conventional models

Analytic model provides direct image position and magnification calculations

Witt--Wynne geometric representation enables quick visual verification

Abstract

We anticipate that hundreds of thousands of distant, strongly gravitationally lensed sources will be detectable with the European Space Agency's (ESA) Euclid mission and the Rubin Observatory Legacy Survey of Space and Time. We consider the virtues and shortcomings of the Singular Isothermal Elliptical Potential (SIEP) with Parallel External Shear (XS_||) for these systems. Its principal virtue is that it admits an analytic forward model that gives image positions and magnifications as functions of the source position (and shape for extended sources). Preliminary experiments suggest a speed-up of a factor in excess of 10,000 compared with conventional models that instead map from the image plane to the source plane and require iteration to converge upon a unique source. A second virtue is that the Witt--Wynne geometric representation of SIEP+XS_|| permits the quick visual verification of the model's adequacy for a particular lensed system. Unfortunately, the model's strictly elliptical lens equipotential is inconsistent with strictly elliptical surface mass density contours. The Witt--Wynne construction might nonetheless yield a sufficiently good first approximation to accelerate convergence to one's preferred lens model.