SITCOMTN-162
Testing the implementation of Metadetection and Cell-Based Coadds on Abell 360 LSSTComCam data#
Abstract
The purpose of this technote is to test the technical quality of LSSTComCam commissioning data, specifically the Rubin_SV_38_7 field, by utilizing cell-based coadds and Metadetection by measuring the tangential and cross weak lensing shear profiles of the massive cluster Abell 360 (called A360 throughout the technote). The process entails generating the cell-based coadds for Metadetection to run on, identifying and removing cluster member galaxies, applying quality cuts, and calibrating the shear measurements. Once a shear profile is generated, validation is the bulk of the remaining analysis.
Cell-based coadds and Metadetection are both currently in the process of being implemented within the LSST Science Pipelines at the time of this technote. There is substantial technical value in attempting a difficult measurement prior to full implementation. Measuring the tangential shear around A360 will showcase the current abilities of these algorithms, as well as highlight where work is still needed.
As seen from the resulting shear profile of A360, the cell-based coadds and Metadetection are able to work in tandem to produce a shear catalog. The shear profile performs best in radial bins further away from the cluster center (beyond ~ 2 Mpc), which may be due to high occurances of blending near the cluster center.
This technote is one part of a series studying A360 in order to both stress test the commissioning camera and demonstrate the technical capabilities of the Vera Rubin Observatory. We study the quality of the PSF modeling and impact it can have on cluster WL in [Combet et al., 2025], implementation of cell-based coadds and subsequent use for Metadetect [Sheldon et al., 2023] in this technote, photometric calibration in (in prep), source selection and photometric redshifts in [Adari and von der Linden, 2025], use of Anacal [Li et al., 2023] to produce a cluster shear profile in [Li, 2025], and background subtraction in this field and Fornax in (in prep).
Cell-Based Coadds Input#
At the time of this analysis, cell-based coadds are not a part of the default LSST Science Pipelines and must be generated independently. The equivalent of the pipetask command below was run on the w_2025_17 weekly stack version of the Pipeline, along with customized branches in drp_tasks and cell_coadds using the branch u/mirarenee/no_ap_corr. The patches and tracts are those that fully or partially fall within 0.5 degrees of the Brightest Cluster Galaxy of A360 at RA, DEC of 37.865017, 6.982205.
The input images and catalogs used to generated the cell-based coadds and other analyses in this technote are from the LSST DRP1 (), focusing on images taken on the Rubin LSSTComCam [SLAC National Accelerator Laboratory and NSF-DOE Vera C. Rubin Observatory, 2024].
pipetask run -j 4 --register-dataset-types \
-b /repo/main \
-i ComCam/runs/DRP/DP1/w_2025_17/DM-50530 \
-o u/$USER/ComCam_Cells/a360 \
-p /sdf/group/rubin/user/mgorsuch/ComCam/pipeline.yaml \
-d "((tract=10463 AND patch IN (30..34,40..44,50..54,60..64,70..74,80..84,90..94)) \
OR (tract=10464 AND patch IN (37..39,47..49,57..59,67..69,77..79,87..89,97..99)) \
OR (tract=10704 AND patch IN (0..5)) \
OR (tract=10705 AND patch IN (8, 9))) \
AND (band='g' OR band='r' OR band='i') AND skymap='lsst_cells_v1'"
The pipeline.yaml file used is shown below:
description: A simple pipeline to test development of cell-based coadds in ComCam
instrument: lsst.obs.lsst.LsstComCam
tasks:
makeDirectWarp:
class: lsst.drp.tasks.make_direct_warp.MakeDirectWarpTask
config:
connections.calexp_list: preliminary_visit_image
connections.visit_summary: visit_summary
connections.warp: direct_warp
connections.masked_fraction_warp: direct_warp_masked_fraction
doWarpMaskedFraction : true
doPreWarpInterpolation : true
makePsfMatchedWarp:
class: lsst.drp.tasks.make_psf_matched_warp.MakePsfMatchedWarpTask
config:
connections.direct_warp: direct_warp
connections.psf_matched_warp: psf_matched_warp
assembleDeepCoadd:
class: lsst.drp.tasks.assemble_coadd.CompareWarpAssembleCoaddTask
config:
connections.inputWarps: direct_warp
connections.psfMatchedWarps: psf_matched_warp
doWriteArtifactMasks : true
assembleCellCoadd:
class: lsst.drp.tasks.assemble_cell_coadd.AssembleCellCoaddTask
config:
connections.inputWarps: direct_warp
connections.visitSummaryList: visit_summary
There are a few reasons why there are additional tasks on top of the cell-based coaddition task, assembleCellCoadd. The primary input images for the cell-based coadds are warped images using the makeDirectWarp task. For cell-based coadds, the doPreWarpInterpolation configuration needs to be set to True manually, as the default pipeline setting is False. This doPreWarpInterpolation config is used to properly propagate the mask plane to the from the warps to the cell-based coadds. The makePsfMatchedWarp and assembleDeepCoadd tasks are needed to generate artifact masks, which are required inputs for the cell-based coaddition task to run.
The collections for the cell-based coadds are stored in u/mgorsuch/a360_cell_coadd for r- and i-bands, and u/mgorsuch/a360_cell_coadd_g for g-band (separate collections due to not running g-band initially).
Cell-based coadds are stored as patch-sized coadds, divided into 484 cell regions (22 by 22 cells). Individual cells have both inner and outer boundary boxes, which are 150 by 150 and 250 by 250 pixels, respectively.
Metadetection utilizes both inner and outer boundaries of cells, and does not duplicate objects from this overlap. However, there is also overlap between patches and tracts that Metadetection does not address, and this leads to duplicate objects within the object tables. Overlap between patches within the same tract is 2 cells wide, and duplicates are removed by removing objects found in the outer ring of cells of each patch. There is also significant overlap between tracts. Patches in tract 10463 that fully overlap with tract 10464 are ignored. After removing the fully overlapping patches, there’s still a 4 cell wide overlap on one side between tract 10463 and 10464. Overlapping cells are again removed. Currently, the WCS information is not enough to remove exact duplicates based on RA and DEC. Why this is the case requires further investigation.
Fig. 1 These three figures show the input image distribution for the patches, each composed of 484 cells, around A360 in the g, r, and i-bands. The red squares outline the inner patch boundaries, where the 2 cell overlap is visible. The three missing patches are due to processing errors when running Metadetection, though they do not significantly overlap with the 0.5 degree radius around the BCG (cyan circle).#
Fig. 2 PSF ellipticity distribution with one PSF realization per cell for the patches around A360 in the g, r, and i-bands. The red squares outline the inner patch boundaries. The general pattern is consistent with [Combet et al., 2025], especially in the r- and i-bands. The mean ellipticities are 0.0563, 0.0689, and 0.0987 for the g, r, and i-bands, respectively.#
Note that for cell-based coadds, there is a single PSF model for each cell, realized at the center of the cell. The distribution of PSF ellipticities are seen in Fig. 2.
Running Metadetection#
Metadetection ([Huff and Mandelbaum, 2017], [Sheldon and Huff, 2017], [Sheldon et al., 2020]) is a shear calibration software focused on an empirical approach of artificially shearing images of galaxies to measure the response calibration matrix R, which is then applied to the unsheared images to calibrate their shear measurements. Metadetection is the sequel software to the original Metacalibration. The primary difference between the two is that while Metacalibration measures the shear response of individual objects for calibration, Metadetection is designed to detect and measure after the applied shear, resulting in 5 catalogs of shear types (non-sheared, in the plus/minus \(g_1\) direction, and in the plus/minus \(g_2\) direction). The main consequence of this is that since detection is shear-dependent, as seen in [Sheldon et al., 2020], the 5 Metadetection catalogs do not have necessarily the same objects, and cannot be matched to each other; shear is instead calibrated using the mean shape values.
The default setting for measuring object shapes is wmom (weighted moments), used throughout this technote. Each shape measurement is a weighted average of the second moments using the three bands, g, r, and i. The weights for averaging across bands come from the inverse variance of the image. In a similar vein, the flux measurements are the zero moment of each object. Both the zero and second moments are weighted by a Gaussian with a FWHM of 1.2 arcseconds.
Metadetection is currently being integrated into the LSST Science Pipelines as a pipeline task to fully utilize the cell-based coaddition based tasks in the pipeline structure.
The Metadetection shear catalog for this technote was run on the w_2025_17 weekly stack version of the Pipeline, along with customized branches in drp_tasks and metadetect using the branch u/mirarenee/meta_test, since a few minor changes were needed to run Metadetection on more recent pipeline stacks.
The pipetask command used to generate the Metadetection catalog is found below:
pipetask run -j 4 --register-dataset-types \
-b /repo/main \
-i refcats,u/mgorsuch/a360_cell_coadd,u/mgorsuch/a360_cell_coadd_g \
-o u/$USER/metadetect/a360_3_band \
-p /sdf/group/rubin/user/mgorsuch/notebooks/metadetect/comcam_pipeline.yaml \
-d "skymap='lsst_cells_v1'"
The associated comcam_pipeline.yaml file used for defining the tasks is outline below:
description: Pipeline for running metadetection on ComCam
instrument: lsst.obs.lsst.LsstComCam
tasks:
metadetectionShear:
class: lsst.drp.tasks.metadetection_shear.MetadetectionShearTask
config:
connections.ref_cat: the_monster_20250219
required_bands : ["g", "r", "i"]
python: |
from metadetect.lsst.configs import get_config as get_mdet_config
mdet_config = get_mdet_config()
mdet_config['metacal']['types']=['noshear', '1p', '1m', '2p', '2m']
config.ref_loader.filterMap = {'lsst_'+band: 'monster_ComCam_%s' % (band) for band in 'ugrizy'}
Currently, with the custom branches, the only python line needed in the pipeline file is the filter map. The custom branches have a workaround in place for changing Metadetection specific configs for the time being, though the additional python lines should be sufficient for changing those configs through the pipeline file in the future.
An important note is that, for this analysis, a single noise image is generated for each cell within Metadetection. The noise image is generated from a Gaussian distribution with a standard deviation equal to the median variance of the image. It’s possible to produce multiple noise images prior to the warping process in order to better account for noise correlation across pixels, though this extra warping is computationally expensive and skipped for now.
Prior to Metadetection flags (i.e. objects cut due to measurement failures), there are 435599 rows after removing duplicate objects due to overlap between patches and tracts. With duplicates removed, there are 81099 objects with Metadetection flags, about 19% of objects. Once all flagged objects are removed (which removes all nan values), there are 354500 remaining objects. Note that this total number is still the combination of all 5 shear type catalogs, with 20% (70902) of the objects being in the non-sheared catalog. The number of flagged objects is quite high, and requires further investigation.
Red Sequence Galaxy Identification#
Cluster member galaxies of the lensing cluster structure will not have a lensing signal (at least not from the cluster itself). Due to this, these galaxies need to be identified and removed from the lensing sample to avoid diluting the shear signal [citation?]. These lensing galaxies are primarily identified through visual inspection using color-magnitude plots across three different bands.
Fig. 3 The distribution of object magnitudes, from Metadetection flux measurements, in each of the bands used. This distribution is after Metadetection flagged objects are cut, though prior to red sequence galaxy identification and selection cuts. The red vertical lines are the current magnitude cuts at the limiting magnitude, determined by eye.#
A series of color-magnitude plots with progressive cuts is used to identify the red sequence (RS) galaxies. Each cut is applied to the entire catalog, though only the non-sheared catalog is shown in the color-magnitude plots. Previous visual inspection showed that while there is some variation in between shear type catalogs, the variation is minimal and random enough that applying the same cuts should be sufficient for this analysis.
The catalog is first cut to galaxies less than 0.1 degree away from the BCG to focus on galaxies that are more likely to be cluster members. The red sequence cluster members are identified in a line of objects with relatively consistent color across a range of magnitudes, with the line being more apparent in the smaller sample of galaxies. This line of galaxies is highlighted with orange points, with the upper and lower limits shown in red. The same visual inspection is done again for the larger sample of galaxies, those within 0.5 degrees of the BCG. Within the larger sample, the objects identified within the limits of either the g-r or r-i RS galaxy limits are cut, leaving a sample of galaxies for measuring weak lensing. This sample still includes bright foreground galaxies, though these should not bias the signal as their shape distribution should be random without the lensing from the cluster. Still, bright galaxies are cut, as seen in the selection cut section.
Fig. 4 Color-magnitude diagram cut to 0.1 degrees within the BCG.#
Fig. 5 Color-magnitude diagram cut to 0.1 degrees within the BCG. Objects that are included within the cut are highlighted in orange.#
Fig. 6 Color-magnitude diagram cut to 0.5 degrees within the BCG.#
Fig. 7 Color-magnitude diagram cut to 0.5 degrees within the BCG. Objects that are included within the cut are highlighted in orange.#
Fig. 8 Color-magnitude diagram cut to 0.5 degrees within the BCG. Objects that are included within the cut are removed.#
After applying the 0.5 degree cut, but prior to removing RS objects, there are 217527 object rows. After applying the RS cuts, there are 152838, with 30532 (20%) of which are the non-sheared catalog.
Selection Cuts#
After Metadetection flags are applied and red sequence galaxies are identified and removed, additional selection cuts are applied. These cuts are primarily based on [Yamamoto et al., 2025], though are made slightly less stringent in order to increase the number of objects used in this relatively small region around the cluster. In numerical terms, the number of objects remaining after selection cuts is 46447, with 9353 (20%) in the non-sheared catalog. Of particular note is the signal-to-noise cut, which doesn’t remove any rows, despite that being a large cut of the total rows in [Yamamoto et al., 2025]. The missing low S/N objects are another quality that requires more investigation.
The Yamamoto cuts describe a size ratio cut, defined as the size of the object squared divided by size of the PSF squared, or \(T^{gauss}/T^{gauss}_{PSF}\). This is used as a star-galaxy cut. For the Yamamoto measurements, these sizes are measured for the pre-PSF objects, so stars will hover around 0 for this ratio. Meanwhile, the weighted moments measurements with Metadetection for T and T_PSF are both measured after the reconvolution step, and will be slightly larger; stars will hover closer to 1. The Metadetection paper [Sheldon et al., 2023] uses a cut of 1.2 for this size ratio, and indicates that while the inclusion of stars might introduce a bias, it’s quite small.
Based on Fig. 11, the stars identified around wmom_T_ratio = 1 appear to fall consistently below wmom_T_ratio = 1.1. This value is chosen instead for the wmom_T_ratio cut, since the inclusion of extra galaxies in the weak lensing sample outweighs the potential inclusion of some low S/N stars.
The magnitude cuts also deviate slightly from [Yamamoto et al., 2025]. These cuts are based on the estimated limiting magnitude of each band, as seen from Fig. 3.
Selection Cut |
Rows Removed |
Fraction Removed |
|---|---|---|
|
59426 |
38.9% |
|
0 |
0.0% |
|
0 |
0.0% |
|
0 |
0.0% |
|
52355 |
34.3% |
|
26595 |
17.4% |
|
37795 |
24.7% |
|
9662 |
6.3% |
|
857 |
0.6% |
|
209 |
0.1% |
|
5352 |
3.5% |
|
181 |
0.1% |
For reference against another catalog, it’s useful to look at the number of objects found in the HSM catalog ([Hirata and Seljak, 2003], [Mandelbaum et al., 2006]) used in [Combet et al., 2025] after different cuts. The HSM catalog first reads in 183791 objects prior to any cuts. After RS cuts, there are 104257 objects. Finally, after selection cuts, the final HSM source galaxy sample is 24362 objects. With the final Metedetection source galaxy sample catalog at 9353 for non-sheared objects, HSM is producing over two times as many objects. This comparison is another sign that the low number of usable objects in the Metadetection catalog needs investigation.
Shear Calibration#
Understanding the relationship between the shear applied to an object and the effect of that shear on measuring the object’s shape is a critical step to calibrating the catalog’s shear measurements. The main purpose of Metadetection’s 5 sub-catalogs is to calculate the linear response matrix, R.
The main assumption is that the weak lensing shear signal is small enough that we can Taylor expand the measured galaxy ellipticity about a zero shear signal. The first term goes to 0 in the limit of a large enough sample of galaxies where the shape noise averages out. The relation between the measured galaxy ellipticity and the resulting shear signal is controlled by R, the linear response of the ellipticity to an applied shear. Within the Metadetection framework, the components of R can be calculated by taking the mean measured galaxy ellipticities of the 4 artificially shear object catalogs. The magnitude of the applied shear in each catalog is 0.01, to make \(\Delta\gamma_j\) a total of 0.02 for each component of R. Once R is calculated, it’s applied to the mean non-sheared object ellipticities to produce the calibrated shear.
In the specific case of cluster lensing, we are more interested in the tangential and cross shears around the cluster, split into radial bins. The response matrix R is first calculated on all galaxy ellipticities measurements from the four sheared catalogs, after all applied cuts, but prior to binning in order to improve the uncertainty of R. Then for each bin, the mean tangential and cross ellipticities are calculated from the uncalibrated galaxy ellipticity measurements in the non-sheared catalog. The response matrix R is then applied to the mean tangential and cross ellipticities to produce the calibrated tangential and cross shears for the bin.
Shear Results#
The resulting shear profile is shown below in Fig. 9. Each bin is calculated from the calibrated shapes of galaxies in six bins that range from 0.32 Mpc to 6.39 Mpc, split evenly in \(\log_{10}\) space. The x-axis shear profile points are calculated from the mean Mpc distance of each object for each bin. The error bars for all shear profiles are bootstrapped samples of each radial bin with 95% confidence levels, which is then calibrated with R. From smallest radial separation to largest, the number of galaxies in each bin are 63, 169, 377, 828, 2028, and 5818 galaxies. The three innermost bins are likely struggling due to the intense blending near the cluster field.
The Mpc distances are assuming a cluster redshift of z=0.22 [Quintana and Ramirez, 1995].
Fig. 9 The reduced shear profile around A360 for both tangential and cross shear measurements, using the cuts described throughout the technote. Both measured profiles have 95% confidence intervals.#
The theoretical shear profile is produced using Cluster Lensing Mass Modeling (CLMM) code [Aguena et al., 2021]. This profile is purely for a rough reference, and is not fit to the calibrated shear data. The profile is using an NFW halo with an estimated cluster mass of 4e14 solar masses ([Hilton et al., 2021]) and a concentration of 4. The source redshift distribution is based off of the DESC Science Requirements Document (SRD, [The LSST Dark Energy Science Collaboration et al., 2018]).
To see if more galaxies were needed in the sample to better characterize R, Metadetection was additionally run on the entire Rubin SV 38 7 field. The errors on the components of R mildly improve, though little difference is seen in the shear profile.
Fig. 10 Shear profile using the same data as Fig. 9, except R is calculated from all galaxies in the Rubin SV 38 7 field produced by Metadetection.#
Galaxies < 0.5 degrees |
Galaxies in SV 38 7 |
|
|---|---|---|
R_11 |
0.2391 |
0.2070 |
R_22 |
0.1796 |
0.2204 |
R_11_err |
0.000935 |
0.0007604 |
R_22_err |
0.000963 |
0.0007670 |
| R_11 - R_22 | |
0.0595 |
0.01336 |
Validation & Testing#
This section contains a non-exhaustive list of notes and figures on characterizing the Metadetection output catalog and the effects from various cuts.
Object Size and S/N Cuts#
The measured object size compared to the signal-to-noise ratio is a simple cut to differentiate stars and galaxies.
Fig. 11 Left: relationship between the object size ratio and the S/N of each object prior to selection cuts, though after red sequence galaxy removal. The red line is a visual reference to see what objects are removed by the 1.1 object size ratio cut. Stars are expected to fall near an object ratio of 1, which is seen clearly for high S/N objects. Right. Distribution of objects after the object size ratio cut (and other additional cuts). The line of stars is removed, though some low S/N stars may survive the cut, as those tend to have higher size uncertainties as seen in [Yamamoto et al., 2025].#
Angular Correlations of PSF Ellipticities#
The version of cell-based coadds used here is not tested on downstream tasks, which is beyond the scope of this technote. Instead, the PSF information from reserved stars (n=212) is taken from the LSSTComCam DRP patch_table using the w_2025_17 weekly pipeline stack version; the specific collection used is LSSTComCam/runs/DRP/DP1/w_2025_17/DM-50530. With this in mind, this section covers some angular correlations of PSF qualities between the reserved stars and the model PSF at the center of each cell. For a more thorough treatment of the PSFs in the Rubin_SV_38_7 field, see [Combet et al., 2025].
The PSF quantities used here have the same definitions for both the reserved star PSFs and the cell-based coadd model PSFs, and both only use the i-band. These are defined with:
The moments used for the reserved stars come from the i_ixx, i_iyy, and i_ixy columns from the object table. As for the model PSF, this is measured directly from the PSF model image for each cell using methods from lsst.afw.math and lsst.meas.algorithms. Angular correlations are calculated using the TreeCorr package ([Jarvis, 2015]), using the default “shot” variance.
Fig. 12 Top row: Residuals of the measured star PSF properties and the cell model PSFs. Note that not all cells will necessarily have a reserved star within them. Bottom row: Angular correlation between residual PSF values. The errors generated from TreeCorr are extremely small and blown up to show relative errors between bins.#
False Detections#
[Yamamoto et al., 2025] introduces a few cuts for spurious detections. The first is an upper limit of wmom_T as a function of wmom_T_err, while the second cut is the product of wmom_T x wmom_T_err. These focus on areas around bright stars and spurious detections within cluster fields, respectively. Plots in Fig. 13 and Fig. 14 were used to get an initial idea of where the data lie. Visual inspection of these objects showed that the majority are around bright stars, spurious detections, and small galaxies with close, undetected neighbors. Still, these cuts remove potentially viable galaxies from the sample, and don’t find every spurious detection. The values used in these cuts come from maximizing the number of spurious detections removed and minimizing the number of viable galaxies kept using visual inspection.
Fig. 13 Distribution of data points before and after cuts. The red lines indicate where the relevant junk cut removes data. The large cluster of points prior to applying all cuts in the left plot is mostly stars removed from the star-galaxy cut, and is not present in the plot on the right.#
Fig. 14 Distribution of data points before and after cuts.#
Object Distributions#
Plotting the distribution of objects on the sky is a simple but effective way to discover inconsistencies within the data. For example, the leftmost plot in Fig. 15 shows an overdensity of objects detected that aligns with where patches overlap, unexpected after removing exact duplicates based on RA and DEC coordinates. This duplication is addressed, as seen in the following plots in Fig. 15.
Fig. 15 Object distributions of the non-sheared catalog at various points during cuts. Left: Galaxy distributions prior to any cuts. There is a clear overdensity that overlaps between patches. Middle: the distribution after the red sequence galaxies have been removed. Right: the object distribution after all cuts have been applied.#
Final Remarks#
From what was able to be achieved within this technote, the combination of cell-based coadds and Metadetection have the necessary infrastructure needed to successfully run end-to end within the LSST Science Pipelines infrastructure and produce a shape catalog. With this shape catalog, a tangential shear profile of A360 was created. Radial bins beyond ~2 Mpc tend to perform the best, while the inner bins struggle to find a clear signal; this is not surprising within a cluster environment.
As for technical details, it should be noted that many configuration settings are not yet available to the pipeline infrastructure through pipeline .yaml files. The main motivation for custom branches was to enable changes (e.g. changing wmom to pgauss) for testing purposes.
There are still plenty of tasks that can done next. It might be beneficial to incorporate alternative deblending algorithms to potentially improve the performance of the shear profile within the inner radial bins near the cluster center. Additionally, running detection tasks on the cell-based coadds would be helpful for comparing objects detected in the Metadetection catalog, as well as obtain PSF measurements for reserved stars in order to do a more thorough \(\rho\)-statistics analysis. More technical investigations should be done to understand why object coordinates differ between patches with the same WCS information, as well as investigate the source of extremely large S/N values and lack of objects with a S/N below 10.
References#
Prakruth Adari and Anja von der Linden. Source Selection for Abell 360 in LSSTComCam Data Preview 1. Commissioning Technical Note SITCOMTN-163, Vera C. Rubin Observatory, July 2025. URL: https://sitcomtn-163.lsst.io/, doi:10.71929/rubin/2571157.
C. Combet, A. Plazas Malagón, S. Fu, P. Adari, I. Dell'Antonio, A. Englert, M. Gorsuch, K. Laliotis, P.-F. Léget, E. Pedersen, A. von der Linden, Y. Zhang, and et al. (TBC). PSF assessment in the field of Abell 360 and shapeHSM shear profile using LSSTComCam data. Commissioning Technical Note SITCOMTN-161, Vera C. Rubin Observatory, July 2025. URL: https://sitcomtn-161.lsst.io/.
Eric Huff and Rachel Mandelbaum. Metacalibration: direct self-calibration of biases in shear measurement. 2017. URL: https://arxiv.org/abs/1702.02600, arXiv:1702.02600.
Xiangchong Li. AnaCal Shear Profile or Abell 360 in ComCam Data. Commissioning Technical Note SITCOMTN-164, Vera C. Rubin Observatory, July 2025. URL: https://sitcomtn-164.lsst.io/.
Xiangchong Li, Rachel Mandelbaum, Mike Jarvis, Yin Li, Andy Park, and Tianqing Zhang. A differentiable perturbation-based weak lensing shear estimator. Monthly Notices of the Royal Astronomical Society, 527(4):10388–10396, December 2023. URL: http://dx.doi.org/10.1093/mnras/stad3895, doi:10.1093/mnras/stad3895.
R. Mandelbaum, C. M. Hirata, M. Ishak, U. Seljak, and J. Brinkmann. Detection of large-scale intrinsic ellipticity–density correlation from the sloan digital sky survey and implications for weak lensing surveys. Monthly Notices of the Royal Astronomical Society, 367(2):611–626, April 2006. URL: http://dx.doi.org/10.1111/j.1365-2966.2005.09946.x, doi:10.1111/j.1365-2966.2005.09946.x.
Erin S. Sheldon, Matthew R. Becker, Michael Jarvis, and Robert Armstrong. Metadetection weak lensing for the vera c. rubin observatory. The Open Journal of Astrophysics, May 2023. URL: http://dx.doi.org/10.21105/astro.2303.03947, doi:10.21105/astro.2303.03947.
Erin S. Sheldon, Matthew R. Becker, Niall MacCrann, and Michael Jarvis. Mitigating shear-dependent object detection biases with metacalibration. The Astrophysical Journal, 902(2):138, oct 2020. URL: https://dx.doi.org/10.3847/1538-4357/abb595, doi:10.3847/1538-4357/abb595.
Erin S. Sheldon and Eric M. Huff. Practical weak-lensing shear measurement with metacalibration. The Astrophysical Journal, 841(1):24, may 2017. URL: https://dx.doi.org/10.3847/1538-4357/aa704b, doi:10.3847/1538-4357/aa704b.
M. Yamamoto, M. R. Becker, E. Sheldon, M. Jarvis, R. A. Gruendl, F. Menanteau, E. S. Rykoff, S. Mau, T. Schutt, M. Gatti, M. A. Troxel, A. Amon, D. Anbajagane, G. M. Bernstein, D. Gruen, E. M. Huff, M. Tabbutt, A. Tong, B. Yanny, T. M. C. Abbott, M. Aguena, A. Alarcon, F. Andrade-Oliveira, K. Bechtol, J. Blazek, D. Brooks, A. Carnero Rosell, J. Carretero, C. Chang, A. Choi, M. Costanzi, M. Crocce, L. N. da Costa, T. M. Davis, J. De Vicente, S. Desai, H. T. Diehl, S. Dodelson, P. Doel, C. Doux, A. Drlica-Wagner, A. Ferté, B. Flaugher, J. Frieman, J. García-Bellido, E. Gaztanaga, G. Giannini, G. Gutierrez, W. G. Hartley, K. Herner, S. R. Hinton, D. L. Hollowood, K. Honscheid, D. Huterer, E. Krause, K. Kuehn, O. Lahav, M. Lima, J. L. Marshall, J. Mena-Fernández, R. Miquel, J. J. Mohr, J. Muir, J. Myles, R. L. C. Ogando, A. Pieres, A. A. Plazas Malagón, A. Porredon, J. Prat, M. Raveri, M. Rodriguez-Monroy, A. Roodman, S. Samuroff, E. Sanchez, D. Sanchez Cid, V. Scarpine, I. Sevilla-Noarbe, M. Smith, E. Suchyta, G. Tarle, V. Vikram, N. Weaverdyck, P. Wiseman, and Y. Zhang. Dark energy survey year 6 results: cell-based coadds and metadetection weak lensing shape catalogue. 2025. URL: https://arxiv.org/abs/2501.05665, arXiv:2501.05665.
M. Aguena, C. Avestruz, C. Combet, S. Fu, R. Herbonnet, A. I. Malz, M. Penna-Lima, M. Ricci, S. D. P. Vitenti, L. Baumont, H. Fan, M. Fong, M. Ho, M. Kirby, C. Payerne, D. Boutigny, B. Lee, B. Liu, T. McClintock, H. Miyatake, C. Sifón, A. von der Linden, H. Wu, M. Yoon, and LSST Dark Energy Science Collaboration. CLMM: a LSST-DESC cluster weak lensing mass modeling library for cosmology. \mnras , 508(4):6092–6110, December 2021. arXiv:2107.10857, doi:10.1093/mnras/stab2764.
M. Hilton, C. Sifón, S. Naess, M. Madhavacheril, M. Oguri, E. Rozo, E. Rykoff, T. M. C. Abbott, S. Adhikari, M. Aguena, S. Aiola, S. Allam, S. Amodeo, A. Amon, J. Annis, B. Ansarinejad, C. Aros-Bunster, J. E. Austermann, S. Avila, D. Bacon, N. Battaglia, J. A. Beall, D. T. Becker, G. M. Bernstein, E. Bertin, T. Bhandarkar, S. Bhargava, J. R. Bond, D. Brooks, D. L. Burke, E. Calabrese, M. Carrasco Kind, J. Carretero, S. K. Choi, A. Choi, C. Conselice, L. N. da Costa, M. Costanzi, D. Crichton, K. T. Crowley, R. Dünner, E. V. Denison, M. J. Devlin, S. R. Dicker, H. T. Diehl, J. P. Dietrich, P. Doel, S. M. Duff, A. J. Duivenvoorden, J. Dunkley, S. Everett, S. Ferraro, I. Ferrero, A. Ferté, B. Flaugher, J. Frieman, P. A. Gallardo, J. García-Bellido, E. Gaztanaga, D. W. Gerdes, P. Giles, J. E. Golec, M. B. Gralla, S. Grandis, D. Gruen, R. A. Gruendl, J. Gschwend, G. Gutierrez, D. Han, W. G. Hartley, M. Hasselfield, J. C. Hill, G. C. Hilton, A. D. Hincks, S. R. Hinton, S. -P. P. Ho, K. Honscheid, B. Hoyle, J. Hubmayr, K. M. Huffenberger, J. P. Hughes, A. T. Jaelani, B. Jain, D. J. James, T. Jeltema, S. Kent, K. Knowles, B. J. Koopman, K. Kuehn, O. Lahav, M. Lima, Y. -T. Lin, M. Lokken, S. I. Loubser, N. MacCrann, M. A. G. Maia, T. A. Marriage, J. Martin, J. McMahon, P. Melchior, F. Menanteau, R. Miquel, H. Miyatake, K. Moodley, R. Morgan, T. Mroczkowski, F. Nati, L. B. Newburgh, M. D. Niemack, A. J. Nishizawa, R. L. C. Ogando, J. Orlowski-Scherer, L. A. Page, A. Palmese, B. Partridge, F. Paz-Chinchón, P. Phakathi, A. A. Plazas, N. C. Robertson, A. K. Romer, A. Carnero Rosell, M. Salatino, E. Sanchez, E. Schaan, A. Schillaci, N. Sehgal, S. Serrano, T. Shin, S. M. Simon, M. Smith, M. Soares-Santos, D. N. Spergel, S. T. Staggs, E. R. Storer, E. Suchyta, M. E. C. Swanson, G. Tarle, D. Thomas, C. To, H. Trac, J. N. Ullom, L. R. Vale, J. Van Lanen, E. M. Vavagiakis, J. De Vicente, R. D. Wilkinson, E. J. Wollack, Z. Xu, and Y. Zhang. The Atacama Cosmology Telescope: A Catalog of >4000 Sunyaev-Zel\textquoteright dovich Galaxy Clusters. \apjs , 253(1):3, March 2021. arXiv:2009.11043, doi:10.3847/1538-4365/abd023.
Christopher Hirata and Uroš Seljak. Shear calibration biases in weak-lensing surveys. \mnras , 343(2):459–480, August 2003. arXiv:astro-ph/0301054, doi:10.1046/j.1365-8711.2003.06683.x.
Mike Jarvis. TreeCorr: Two-point correlation functions. Astrophysics Source Code Library, record ascl:1508.007, August 2015.
H. Quintana and A. Ramirez. Redshifts of 165 Abell and Southern Rich Clusters of Galaxies. \apjs , 96:343, February 1995. doi:10.1086/192122.
SLAC National Accelerator Laboratory and NSF-DOE Vera C. Rubin Observatory. Lsst commissioning camera. 2024. URL: https://www.osti.gov//servlets/purl/2561361, doi:10.71929/RUBIN/2561361.
The LSST Dark Energy Science Collaboration, Rachel Mandelbaum, Tim Eifler, Renée Hložek, Thomas Collett, Eric Gawiser, Daniel Scolnic, David Alonso, Humna Awan, Rahul Biswas, Jonathan Blazek, Patricia Burchat, Nora Elisa Chisari, Ian Dell'Antonio, Seth Digel, Josh Frieman, Daniel A. Goldstein, Isobel Hook, Željko Ivezić, Steven M. Kahn, Sowmya Kamath, David Kirkby, Thomas Kitching, Elisabeth Krause, Pierre-François Leget, Philip J. Marshall, Joshua Meyers, Hironao Miyatake, Jeffrey A. Newman, Robert Nichol, Eli Rykoff, F. Javier Sanchez, Anže Slosar, Mark Sullivan, and M. A. Troxel. The LSST Dark Energy Science Collaboration (DESC) Science Requirements Document. arXiv e-prints, pages arXiv:1809.01669, September 2018. arXiv:1809.01669, doi:10.48550/arXiv.1809.01669.