An Image-Plane Approach to Gravitational Lens Modeling of Interferometric Data

TL;DR

This paper introduces an efficient image-plane lens modeling method for interferometric data, validated with simulations and real ALMA observations, offering comparable accuracy to traditional visibility-based approaches.

Contribution

The authors develop and implement an image-plane modeling technique for interferometric gravitational lensing data, reducing computational costs while maintaining accuracy.

Findings

Image-plane likelihood yields accurate models with noise correlations.

Method performs well on simulated and real ALMA data.

Results are consistent with previous visibility-based models.

Abstract

Strong gravitational lensing acts as a cosmic telescope, enabling the study of the high-redshift universe. Astronomical interferometers, such as the Atacama Large Millimeter/submillimeter Array (ALMA), have provided high-resolution images of strongly lensed sources at millimeter and submillimeter wavelengths. To model the mass and light distributions of lensing and source galaxies from strongly lensed images, strong lens modeling for interferometric observations is conventionally performed in the visibility space, which is computationally expensive. In this paper, we implement an image-plane lens modeling methodology for interferometric dirty images by accounting for noise correlations. We show that the image-plane likelihood function produces accurate model values when tested on simulated ALMA observations with an ensemble of noise realizations. We also apply our technique to ALMA observations of two sources selected from the South Pole Telescope survey, comparing our results with previous visibility-based models. Our model results are consistent with previous models for both parametric and pixelated source-plane reconstructions. We implement this methodology for interferometric lens modeling in the open-source software package lenstronomy.