SourceREACH (Source REconstruction of Arcs behind Cluster Halos): A New Source Reconstruction Algorithm Optimized for Giant Arcs and Galaxy Cluster Lenses
Lana Eid, Charles R. Keeton · 2025
Read SourceREACH (Source REconstruction of Arcs behind Cluster Halos): A New Source Reconstruction Algorithm Optimized for Giant Arcs and Galaxy Cluster Lenses by Lana Eid, Charles R. Keeton (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-11952.txt.
Introduction
Lana Eid and Charles Keeton introduce SourceREACH, an algorithm that reconstructs the true appearance of distant galaxies whose light is stretched into giant arcs by a foreground galaxy cluster's gravity. Cluster lenses magnify the early universe, but modeling high-resolution arcs is computationally punishing.
The method deconvolves the telescope's blur, de-lenses each pixel back to the source plane, and builds the source by interpolation or smoothed regression. Working with de-lensed points handles the wildly varying effective resolution across an arc. Speed with accuracy matters because giant arcs carry joint information about dark matter in the cluster and star formation in the source, and upcoming surveys will deliver them in bulk.
Catalog record
| CREATOR | Lana Eid, Charles R. Keeton |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-11952
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
SourceREACH (Source REconstruction of Arcs behind Cluster Halos): A New Source Reconstruction Algorithm Optimized for Giant Arcs and Galaxy Cluster Lenses Lana Eid, Charles R. Keeton Abstract We introduce a new algorithm designed for use with extended lensed images, specifically giant arcs lensed by galaxy clusters. These highly magnified images contain important information about both the mass distribution of the cluster and the properties of the background source, but modeling them requires significant computational effort. Our new source reconstruction methodology is designed to be accurate and efficient for high-resolution observations in which point spread function effects are not significant. The overall process deconvolves the observed image by the point spread function, de-lenses the image pixels, and uses interpolation or regression with smoothing to determine the model source. By working with de-lensed points, the method accounts for varying resolution across the source plane. We evaluate the speed and accuracy of different interpolation and regression methods using both mock data and real data for the giant arc in Abell 370. We find that utilizing K Nearest Neighbor Regression results in the best balance of noise smoothing and preservation of compact detail in the source. I Introduction Galaxy cluster lens models inform us about the cluster-scale mass distribution ( 122 ; 31 ; 90 , e.g.,) , possible substructure on a smaller scale ( 38 , e.g.,) , and cosmological parameters like the Hubble constant and dark energy from time delays ( 152 ; 22 ; 56 ; 116 , e.g.,) . Corresponding source reconstruction models facilitate spatially resolved studies of overall source properties, including star formation rate, locations of emission line regions such as Lyman \alpha , and investigations of the smallest scales on which star formation occurs ( 125 ; 63 ; 112 ; 25 ; 103 , e.g.,) . Giant, spatially extended arcs crossing critical curves are highly-magnified, multiply lensed extended objects, so they offer us maximum insight into both the source in the early Universe as well as regions producing these images. Studies of both high-redshift source properties ( 63 ; 140 ; 25 ; 39 ; 50 ; 103 ) and the small-scale lensing dark matter structure effects of milli-lensing ( 104 ; 40 ; 50 ; 90 , e.g.,) rely on the accuracy of multiple-image source reconstructions and the highest magnification regions marked by corresponding critical curves. Microlensing events at critical curves ( 38 ; 40 , e.g.,) can contain even higher magnifications than giant arcs overall, and constraining cluster critical curves is crucial for comparing to and interpreting such events. Recent results, particularly from the Hubble Frontier Fields multi-team modeling efforts ( 121 ; 163 ; 44 , e.g.,) , have demonstrated that these highest magnification regions are still uncertain, with wide variation and systematic uncertainty even among models that generally fit the lensed point-images equally well ( 86 ; 53 ; 52 ; 51 ; 89 ; 90 ) . It is therefore essential to develop a methodology to best constrain these highest magnification regions of cluster models with high-resolution data for giant arcs from the wealth of information we have now and in upcoming targeted observations and large surveys ( 138 ) . Source reconstruction can be done in many forms, including semi-linear inversion, pixel-based, and ray-tracing ( 159 ; 146 ; 154 ; 148 ; 147 ; 105 ; 37 ; 150 ; 71 ; 164 ; 136 ; 132 , e.g.,) , forward-modeling with iteratively added complexity ( 163 , e.g.,) , and back-projection using a modeled magnification matrix ( 57 , e.g.,) . Bayesian probabilistic methods are often built into these methods to incorporate uncertainties and regularization; varying source plane resolution can be handled using adaptive gridding techniques, including Delaunay triangulation and Voronoi tessellation. Techniques for constraining both the lens and background source simultaneously are computationally expensive for cluster-scale models; approaches such as in 9 ; 37 ; 10 ; 163 ; 132 ; 1 were tested on separate, multiply lensed, extended images, while the highest magnification regions of interest, which giant arcs cross multiple times, are comprised of merging images that are more computationally expensive to handle as a cluster model constraint. Recent studies have used lensed arcs to constrain local regions of cluster mass models ( 11 ; 91 ; 136 ; 137 , e.g.,) . However, the resulting local mass model may not be not valid for the entire cluster due to the focus on more accurately constraining just that region. 91 show that matching different regions across extended images can adjust local regions of a cluster model without affecting the global properties of the cluster model. We similarly showed this for the case of the giant arc in Abell 370 ( 44 , hereafter Paper 1) , in which we locally optimized a fiducial cluster model from 122 using full a pixel-based source reconstruction without significantly affecting the rest of the lensed image position fits or adjusting the cluster-scale matter halos. Ideally, the entire cluster model would be optimized at once to ensure the best fit for the entire field. We aim to facilitate this with a resource-efficient source reconstruction methodology that we term SourceREACH (Source REconstruction of Arcs behind Cluster Halos) to work with any set of standard output maps, such as from the Hubble Frontier Fields ( 93 ; 163 ) . We perform a comprehensive study comparing different interpolation methods for an updated de-lensing python code, prototyped in Paper 1, and we validate our final choice using Hubble Frontier Fields optical data ( 93 ) . Section II presents the data and mock data used, the code development, and the testing; Section III presents our results; Section IV interprets the results; and Section V contextualizes them for future studies. II Data & Methods: II.1 The Abell 370 Galaxy Cluster Field We use data from the Hubble Frontier Fields (HFF) Program ( 93 ) , which was a 3-year, 840-orbit Hubble Space Telescope (HST) targeting six galaxy cluster fields with the aim of obtaining deep images of the clusters and faint z\sim 5-10 background galaxies magnified by strong lensing. This survey program is ideal for our study, since it is the base of the various comparison studies that have followed from the lens modeling teams that used the same data to construct their cluster models ( 121 ; 163 , e.g.,) . We focus on the z=0.375 Abell 370 cluster field, which includes the z=0.725 first confirmed strongly lensed galaxy in a giant arc ( 145 ; 143 ; 144 ) . We chose this well-studied system of an extended, giant arc comprised of five merging images of the same background galaxy for our in-depth study of how a high-resolution source reconstruction can be efficiently leveraged as a constraint in order to help address systematic differences between results from diverse cluster modeling methods that persist in the predicted high-magnification regions of interest in these cluster models ( 121 ; 52 ) . Paper 1 details the steps we took to isolate the arc and format our inputs for each filter and corresponding data image; we briefly outline them here. The top left panel of Figure 1 shows the data image from the F814W filter, which