Pith. sign in

REVIEW 3 major objections 4 minor 51 references

Direct wave simulations of fuzzy dark matter halos can reproduce the observed image positions of the quadruply lensed quasar HS 0810+2554 to within about 3 milliarcseconds, matching the data better than Gaussian-random-field or smooth-halo

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Wave-simulated fuzzy dark matter halos at 10^-22 eV produce lensed-image position anomalies for HS 0810+2554 that match observations better than Gaussian-random-field or NFW models.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection First direct Schrödinger–Poisson FDM lensing predictions, but the headline 3σ match is an uncalibrated best-of-selection; the method is worth referee time with revisions. the 3 major comments →

arxiv 2607.20614 v1 pith:2GJHGCLM submitted 2026-07-22 astro-ph.CO

Gravitational Lensing Predictions from Wave Simulations of Fuzzy Dark Matter

Jiajun Zhou , Zhengxiang Li , Amruth Alfred , Ran Gao , Jeremy Lim , Zong-Hong Zhu This is my paper
classification astro-ph.CO
keywords fuzzy dark matterSchrödinger–Poissongravitational lensingastrometric anomaliesHS 0810+2554wave interferenceGaussian random fieldhalo simulations
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that the granular, wave-like density fluctuations predicted in fuzzy dark matter halos—computed by directly evolving the Schrödinger–Poisson equations rather than approximating them as Gaussian random fields—can explain the anomalously displaced image positions of the quadruply lensed quasar HS 0810+2554. For a dark matter particle mass of 10^-22 eV, the simulated halos give median position anomalies of ~12 and ~6 milliarcseconds for the two radio-loud image components, and a subset of realizations falls within the ~3 milliarcsecond (3σ) astrometric uncertainty. The authors emphasize that this is achieved without any best-fit tuning of the halo profile, in contrast to the best-fit NFW model and the GRF approximation, which leave larger residual anomalies. If the claim holds, high-resolution imaging of lensed quasars becomes a direct probe of the wave nature of dark matter and of the particle mass.

Core claim

The central discovery is that lensing predictions made from 1000 Schrödinger–Poisson evolved halos of mass 4×10^11 M⊙ produce a distribution of image-position anomalies whose best realizations are small enough to sit inside the observed astrometric uncertainty of HS 0810+2554, at a level the authors state is better than both the Gaussian-random-field approximation and a best-fit NFW model. The evolved halos naturally contain a central soliton core surrounded by granular interference fluctuations with ~10% projected-density contrast; these fluctuations perturb the critical curve and shift the lensed images. For mψ = 10^-22 eV the median anomaly is 12 and 6 mas for the two radio components, an

What carries the argument

The load-bearing machinery is the direct numerical solution of the Schrödinger–Poisson system with a global Fourier pseudospectral split-step method, which yields 1000 independent three-dimensional halos from random multi-wave-packet initial conditions. Each halo is projected along 10^3 sampled viewing directions, and for each projection the lens equation is solved for 10^4 source positions; the relative geometry of the resulting four-image configuration is compared with the observed radio images using the six pairwise image separations and a Procrustes shape-matching step that removes global translation and rotation. This forward-modelling search identifies the viewing direction and source

Load-bearing premise

The entire comparison rests on the assumption that selecting the best matching viewing direction and source position from 1000 halos (with 1000 orientations and 10,000 source positions each) does not itself generate sub-3-milliarcsecond image offsets purely by chance; if that search freedom is not accounted for, the claimed match to HS 0810+2554 is not yet evidence for 10^-22 eV fuzzy dark matter.

What would settle it

Run the same forward-modelling search—1000 realizations, 1000 viewing directions, 10,000 source positions—on smooth NFW halos with no wave fluctuations, and count the fraction of realizations with A<3 mas for the HS 0810+2554 geometry. If smooth halos yield a comparable fraction of sub-3-mas 'matches', the wave-simulated result is an artifact of the search. Alternatively, demand that a single halo, with a single orientation and source position, reproduce all eight lensed image positions simultaneously within 3σ; the paper's component-wise 'some realizations' claim would fail this stricter test

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If fuzzy dark matter halos at 10^-22 eV really produce sub-3-mas image offsets, then astrometric measurements of multiply lensed quasars can be used to pin down the dark matter particle mass, since the predicted anomaly scales strongly with mass (10^-23 eV gives roughly fourfold larger offsets).
  • Because the wave-simulated halos outperform the Gaussian-random-field approximation, inferences about dark matter drawn from GRF-based lensing models will need to be re-checked with full wave evolution; the GRF may remain adequate for average statistics but not for individual systems.
  • The result suggests that the observed position anomalies in HS 0810+2554 do not require additional lens-model complexity such as multipole perturbations or external shear; the wave fluctuations in a single FDM halo can supply the needed perturbative power.
  • The same forward-modelling framework can be extended to flux-ratio anomalies, time delays, and resolved image morphologies, which the authors list as planned next steps and which would provide independent tests of the same fuzzy-dark-matter hypothesis.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The paper's 'no best-fitting procedure' statement is only partly fair: the pipeline still searches over 10^3 orientations and 10^4 source positions per halo and reports the best match, so the 3-mas realizations are selected from a large ensemble. A sharper test would be to run the identical search on smooth NFW halos or pure noise and compare how often A<3 mas arises by chance; until then, part of
  • If the best-of-1000 selection is not the main driver, the implication is that the match is statistical—most FDM halos do not fit, but a non-negligible fraction do—so a single lens system cannot by itself confirm 10^-22 eV dark matter; the natural next step is to apply the same pipeline to a sample of lens systems and ask whether the observed anomalies are drawn from the predicted distribution.
  • The appendix's Gaussianity tests show that the projected S-P fluctuations are nearly Gaussian at fixed radius, which means the GRF approximation is not wrong in a statistical-average sense; what full wave evolution adds is the correct spatial correlation structure and non-Gaussian tail behavior, which matter for the specific image configurations of individual lenses. A practical extension is to bu
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. This paper presents gravitational lensing predictions generated from 3D density fields obtained by Schrödinger–Poisson (SP) simulations of fuzzy dark matter (FDM) halos. For m_ψ = 10^-22 eV, the authors simulate 1000 halo realizations, project each along optimized viewing directions, and forward-model lensed image positions using the public lenstronomy package. They compare the position anomaly A (Eq. 2) with the two radio components of HS 0810+2554, reporting median anomalies of ~12 and ~6 mas, with some realizations within ~3 mas (3σ). They also compare against GRF-based approximations and a smooth NFW model, claiming the SP halos reproduce the observed image positions better than both. The central claim is that wave-simulated halos can reproduce the observed positions 'despite involving no best-fitting procedure.'

Significance. The framework is a potentially important methodological step beyond the GRF approximation used in earlier FDM lensing papers, and the direct use of SP evolution is physically better motivated. The authors make a concrete, falsifiable prediction and use a widely available lensing code; the three-model comparison on the same system is a useful sanity check. However, the headline '3σ match' is not yet supported because the reported anomaly is the minimum over a very large parameter search (§6.2), and the claim of no fitting is contradicted by the pipeline. With proper null calibration and a more careful statement of what is being tested, this could become a valuable contribution.

major comments (3)
  1. [§6.2, Eqs. (20)–(23)] The 'best' anomaly for each halo is the minimum over N_dir=10^3 viewing directions (Eq. 20) and N_src=10^4 source positions (Eq. 23), and Figure 4 shows only the 300 smallest-A realizations out of 1000. This is effectively a best-of-10^10 selection. The paper provides no null calibration (e.g., the same search applied to a smooth NFW lens or to randomly permuted observed image positions), so the expected minimum A under the null is unknown. The statement in §3 that 'some realizations are within 3σ' is an uncalibrated extreme-value statistic. This is load-bearing because it underpins the conclusion that SP halos match the observations.
  2. [§3, final paragraph] The claim that the match arises 'despite involving no best-fitting procedure' is contradicted by §6.2, where the viewing direction (Eq. 20) and source position (Eq. 23) are free parameters chosen to minimize A, followed by Procrustes alignment (Eqs. 24–28) that removes translation, rotation, and reflection. The same paragraph also states that images from different realizations can be combined to reproduce all eight lensed images, which implies that no single physical halo actually matches all eight. The paper should be reframed as a statistical distribution of anomalies, not as a demonstration that one wave-simulated halo reproduces the system.
  3. [§2 and §6.1] The halo ensemble is generated from five initially random Gaussian wave packets in an isolated 40 kpc box with fixed total mass M = 4×10^11 M⊙. This is not a cosmological simulation, and it is unclear whether it samples the diversity of real FDM halos (triaxiality, soliton position, merger history, concentration). If the ensemble is biased, the fact that some realizations match after a massive orientation/source search is less informative. The authors should demonstrate that the statistical properties of the projected convergence fields are consistent with cosmological FDM simulations, or explicitly discuss this limitation.
minor comments (4)
  1. [Eq. (2)] The expression for A is written as |(x_obs,i - x_pred,i) + (y_obs,i - y_pred,i)|^2, which is not the standard Euclidean distance squared. It should be (Δx)^2 + (Δy)^2. Please correct the notation.
  2. [Fig. 4 caption] The caption states 'only the 300 best-matching realizations' are shown, but the text says '300 unique FDM halos.' It should be clarified whether the selection is made per radio component or jointly, and what criterion defines 'best.'
  3. [§6.5] The description refers to 'Equation (8) of A. Amruth et al. (2023)' without reproducing the equation. For self-containedness, include the relevant equation or a complete description of how σ_κ is computed.
  4. [Abstract and §1] The text refers to 'the quadruply-lensed radio jets in system HS 0810+2554.' The system has two radio components, each quadruply lensed (eight images). The wording should be adjusted to avoid ambiguity.

Circularity Check

2 steps flagged

The '3σ, no best-fitting' headline is the minimum of the paper's own orientation/source-position optimization and a halo mass chosen to match the Einstein radius, so the headline prediction reduces to the fitted residual.

specific steps
  1. fitted input called prediction [Section 2 (simulation setup) and Section 3 (summary claim)]
    ""These parameter choices are determined only by the approximate halo-mass scale of the lens system with the aim of reproducing the correct Einstein radius and the observed lensing configuration (See Appendix for further discussion). ... It is remarkable that the wave-simulated halos can reproduce the observed image positions of HS 0810+2554 to within 3σ, despite involving no best-fitting procedure: the halos are evolved from initial conditions without any control over their final states.""

    The total halo mass M=4×10^11 M⊙ is selected "with the aim of reproducing the correct Einstein radius and the observed lensing configuration" of the target system; the Einstein radius is set by the observed image positions. Calling the subsequent agreement "no best-fitting procedure" ignores that a global parameter was already fitted to the target system. The halo evolution is independent only conditional on that mass choice.

  2. fitted input called prediction [Appendix §6.2, Eqs. (20)–(29); Section 3]
    ""For each principal-axis direction, the source position that minimizes this norm is selected as the optimal source position. ... By comparing the residual anomaly A over all sampled directions, we identify the best-matching pair (ê_k^best, β_m^best), which specifies the preferred spatial orientation of the simulated FDM density field and the source position that most closely reproduce the observed relative image geometry.""

    By construction, the reported A is the minimum of Eq. (29) (Procrustes-aligned rms) over 10^3 viewing directions (Eq. 20) and 10^4 source positions (Eq. 23). The histograms in Fig. 3 and the "some realizations within 3σ" statement are therefore best-fit residuals, not predictions from a fixed model with no free parameters. The observed positions are the objective of the search; the "no best-fitting" claim is contradicted by the paper's own matching procedure.

full rationale

The Schrödinger–Poisson simulations themselves are not circular: they are evolved independently and compared against GRF/NFW. The relative ordering wave-simulated vs GRF may survive because both are processed the same way. However, the absolute headline claim — 10^-22 eV FDM halos reproduce HS 0810+2554 within 3σ with no fitting — is materially weakened because (i) the halo mass is set by the requirement to reproduce the Einstein radius / observed lensing configuration, and (ii) the 3 mas value is the minimum over 10^3×10^4 choices of orientation and source position, plus Procrustes alignment. The paper itself describes the search for 'the source position that minimizes this norm' and 'the best-matching pair'. Thus the 'predicted' anomaly reduces, by construction, to a fitted residual. This is a partial circularity: the central claim as stated is not supported as a parameter-free prediction, although the pipeline has independent components. No load-bearing self-citation or uniqueness-import issue found; self-citations to Amruth et al. (2023) supply the GRF prescription and are not used to force the main result.

Axiom & Free-Parameter Ledger

6 free parameters · 6 axioms · 0 invented entities

The central result rests on the assumption that evolved isolated halos initialized from five Gaussian wave packets represent realistic FDM halos, and on optimized nuisance parameters (orientation and source position) that are fitted to the observed image positions. The particle mass and total halo mass are model inputs rather than numbers derived from the lens data. No new entities are introduced.

free parameters (6)
  • FDM particle mass mψ = 10^-22 eV (and 10^-23 eV for comparison)
    Model input scanned, not fit to lens data; the central claim depends on this value.
  • Total halo mass M = 4×10^11 M⊙
    Chosen to match the approximate halo-mass scale and Einstein radius of HS 0810+2554 (§2).
  • Box size L / grid N = 40 kpc / 512
    Resolution choice; together with mψ sets the de Broglie wavelength and fluctuation properties.
  • Initial condition: number of Gaussian wave packets = 5
    Ad hoc modeling choice; no physical justification for exactly five wave packets, and it affects final halo structure.
  • Viewing direction θ per halo = Optimized per realization over 10^3 directions
    Nuisance parameter fitted to observed image geometry (§6.2); selection can reduce A artificially.
  • Source position (βx, βy) per halo = Optimized per realization over 10^4 sources
    Nuisance parameters fitted to observed image geometry (§6.2); standard in lens modeling but part of the matching.
axioms (6)
  • domain assumption Schrödinger–Poisson equations govern FDM halo dynamics
    Standard model for FDM; invoked in Eq. 3 and throughout the simulation.
  • domain assumption Thin-lens approximation is valid for projecting 3-D density to convergence
    Used in Eq. 21 and throughout §3; standard for galaxy-scale lensing but ignores 3-D lensing effects.
  • domain assumption Planck 2020 cosmology and redshifts z_s=1.51, z_l=0.89 for HS 0810+2554
    Used to set the evolution-time upper bound and angular scale (§6.1).
  • ad hoc to paper Isolated-box evolution with five wave packets yields statistically representative lens halos
    No cosmological infall, tidal field, or baryons are included; the assumption is load-bearing for the comparison and is not validated against full cosmological FDM simulations.
  • domain assumption Observed astrometric uncertainties from Hartley et al. 2019 are correct
    The 3σ thresholds and the anomaly scale used throughout the paper depend on these published uncertainties.
  • domain assumption GRF construction of Amruth et al. 2023 is the appropriate baseline
    The GRF comparison is built on their prescription; this is a reasonable prior baseline, but it is not the only possible smooth+fluctuation model.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Gravitational Lensing Predictions from Wave Simulations of Fuzzy Dark Matter." pith.science (2026). https://pith.science/paper/2GJHGCLM

@misc{pith2026260720614,
  author       = {Pith},
  title        = {Pith review of: Gravitational Lensing Predictions from Wave Simulations of Fuzzy Dark Matter},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2GJHGCLM}},
  note         = {Machine review of arXiv:2607.20614}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

In the cold dark matter paradigm, ultra-light particles are emerging as strong contenders to conventional massive particles. A unique prediction of dark matter comprising such ultra-light particles, known as fuzzy dark matter (FDM), is the presence of strong density modulations throughout galactic halos due to wave interference, which -- when approximated by a Gaussian random field (GRF) -- have been proposed to account for the inability to reproduce the observed positions (when measured at sufficient precisions) and flux ratios of multiply-lensed images of quasars. Here, we predict for the first time the properties of gravitationally lensed images generated from 3-D density fields obtained by wave simulations that directly evolve the Schr\"odinger--Poisson equations. Using a novel framework to project these evolved density fields along various axes of the 3-D halo, we obtain the distribution of perturbations to the positions of lensed images. As an exacting test, we find that particles of mass $10^{-22}$ eV can reproduce the positions of the quadruply-lensed radio jets in system HS 0810+2554 to a level better than that of either the GRF approximation or, to a greater extent, an NFW best-fit solution, both of which rely on accurately capturing the global 3-D density field of dark matter halos. Our work highlights the importance of wave simulations for making accurate FDM lensing predictions and the potential for high-resolution observations of lensed systems to serve as a direct probe of the nature of dark matter.

Figures

Figures reproduced from arXiv: 2607.20614 by Amruth Alfred, Jeremy Lim, Jiajun Zhou, Ran Gao, Zhengxiang Li, Zong-Hong Zhu.

Figure 1
Figure 1. Figure 1: Projected density field of an evolved fuzzy dark matter halo obtained from the Schr¨odinger–Poisson simulation (mψ = 10−22 eV). From left to right, the panels show the convergence κFDM (projected surface mass density relative to the critical density), its elliptical-annular mean ⟨κFDM⟩, and the relative residual between the total and mean convergence (κFDM − ⟨κFDM⟩)/⟨κFDM⟩. The ellipticity and position ang… view at source ↗
Figure 2
Figure 2. Figure 2: Spherically averaged radial density profiles of the evolved fuzzy dark matter halos. The blue and orange curves correspond to particle masses of mψ = 10−22 eV and mψ = 10−23 eV, respectively. The blue and orange dotted curves show the corresponding best-fitting soliton profiles in the inner halo regions (H.-Y. Schive et al. 2014b). The black dashed and solid curves show the best-fitting NFW profiles for th… view at source ↗
Figure 3
Figure 3. Figure 3: Probability distributions of the image-posi￾tion anomaly A. (a) Schr¨odinger–Poisson FDM simulations with mψ = 10−22 eV. (b) GRF-based FDM density realiza￾tions constructed on top of the best-fitting NFW model. (c) Schr¨odinger–Poisson FDM simulations with mψ = 10−23 eV. The blue and red histograms show the normalized distribu￾tions for the two radio sources, radio1 and radio2, respec￾tively, while the sol… view at source ↗
Figure 4
Figure 4. Figure 4: Predicted and observed lensed image positions for the two radio components of HS 0810+2554. The central panel shows the full image configuration, while the surrounding panels zoom into the individual images. To avoid visual overcrowding, only the 300 best-matching realizations with the smallest position anomaly A are shown, with the images denoted by colored points. The black solid ellipses denote the 3σ m… view at source ↗
Figure 5
Figure 5. Figure 5: Schematic illustration of the reduced matching procedure. (a) The line of sight is fixed along the z-axis, while the principal axis of the three-dimensional density field is rotated to representative directions in the positive xz plane. (b) Four-image configurations related by trans￾lation, rotation, or reflection are treated as geometrically equivalent. In the actual calculation, all six pairwise dis￾tanc… view at source ↗
Figure 6
Figure 6. Figure 6: Annular rms amplitudes of the projected convergence fluctuations for the Schr¨odinger–Poisson (S–P) and Gaussian random field (GRF) realizations. The left and right panels show the fractional rms, ⟨(δκ/κref) 2 ⟩ 1/2 , and the absolute rms, ⟨(δκ) 2 ⟩ 1/2 , respectively. Here, δκ = κ − κref, with κref = ⟨κFDM⟩ell for the S–P realization and κref = κNFW for the GRF realization. The vertical dashed lines mark … view at source ↗
Figure 7
Figure 7. Figure 7: Gaussianity tests for the standardized residual fluctuations in three representative annuli. To match the adopted de Broglie wavelength, λdB ∼ 120 pc, each annulus is assigned a radial width of ∆R = 120 pc. For each annulus, the left panel compares the probability-density distributions of the standardized residuals from the fuzzy dark matter (FDM) and Gaussian random field (GRF) realizations with the stand… view at source ↗
Figure 8
Figure 8. Figure 8: Extension of [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Extension of [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 1 linked inside Pith

  1. [1]

    2023, Nature Astronomy, 7, 736, doi: 10.1038/s41550-023-01943-9

    Amruth, A., Broadhurst, T., Lim, J., et al. 2023, Nature Astronomy, 7, 736, doi: 10.1038/s41550-023-01943-9

    doi:10.1038/s41550-023-01943-9 2023

  2. [2]

    Banik, N., Bovy, J., Bertone, G., Erkal, D., & de Boer, T. J. L. 2021, JCAP, 2021, 043, doi: 10.1088/1475-7516/2021/10/043

    doi:10.1088/1475-7516/2021/10/043 2021

  3. [3]

    2018, Physics of the Dark Universe, 22, 189, doi: 10.1016/j.dark.2018.11.002

    Birrer, S., & Amara, A. 2018, Physics of the Dark Universe, 22, 189, doi: 10.1016/j.dark.2018.11.002

    doi:10.1016/j.dark.2018.11.002 2018

  4. [4]

    J., Gilman, D., et al

    Birrer, S., Shajib, A. J., Gilman, D., et al. 2021, Journal of Open Source Software, 6, 3283, doi: 10.21105/joss.03283

    doi:10.21105/joss.03283 2021

  5. [5]

    S., & Kaplinghat, M

    Boylan-Kolchin, M., Bullock, J. S., & Kaplinghat, M. 2011, MNRAS, 415, L40, doi: 10.1111/j.1745-3933.2011.01074.x

    arXiv 2011

  6. [6]

    S., & Kaplinghat, M

    Boylan-Kolchin, M., Bullock, J. S., & Kaplinghat, M. 2012, MNRAS, 422, 1203, doi: 10.1111/j.1365-2966.2012.20695.x

    arXiv 2012

  7. [7]

    N., Smoot, G

    Broadhurst, T., De Martino, I., Luu, H. N., Smoot, G. F., & Tye, S.-H. H. 2020, Phys. Rev. D, 101, 083012, doi: 10.1103/PhysRevD.101.083012

    doi:10.1103/physrevd.101.083012 2020

  8. [8]

    S., & Boylan-Kolchin, M

    Bullock, J. S., & Boylan-Kolchin, M. 2017, ARA&A, 55, 343, doi: 10.1146/annurev-astro-091916-055313

    doi:10.1146/annurev-astro-091916-055313 2017

  9. [9]

    2020, PhRvL, 125, 111102, doi: 10.1103/PhysRevLett.125.111102

    Broadhurst, T. 2020, PhRvL, 125, 111102, doi: 10.1103/PhysRevLett.125.111102

    doi:10.1103/physrevlett.125.111102 2020

  10. [10]

    2017, MNRAS, 468, 1338, doi: 10.1093/mnras/stx449

    Chen, S.-R., Schive, H.-Y., & Chiueh, T. 2017, MNRAS, 468, 1338, doi: 10.1093/mnras/stx449

    doi:10.1093/mnras/stx449 2017

  11. [11]

    Dalal, N., & Kochanek, C. S. 2002, ApJ, 572, 25, doi: 10.1086/340303

    doi:10.1086/340303 2002

  12. [12]

    2022, PhRvD, 106, 063517, doi: 10.1103/PhysRevD.106.063517 de Blok, W

    Dalal, N., & Kravtsov, A. 2022, PhRvD, 106, 063517, doi: 10.1103/PhysRevD.106.063517 de Blok, W. J. G. 2010, Advances in Astronomy, 2010, 789293, doi: 10.1155/2010/789293

    doi:10.1103/physrevd.106.063517 2022

  13. [13]

    W., & Witt, H

    Evans, N. W., & Witt, H. J. 2003, MNRAS, 345, 1351, doi: 10.1046/j.1365-2966.2003.07057.x

    arXiv 2003

  14. [14]

    Ferreira, E. G. M. 2021, A&A Rv, 29, 7, doi: 10.1007/s00159-021-00135-6

    doi:10.1007/s00159-021-00135-6 2021

  15. [15]

    2025, PhRvD, 111, L081302, doi: 10.1103/PhysRevD.111.L081302

    Gao, R., Tian, S., Li, Z., et al. 2025, PhRvD, 111, L081302, doi: 10.1103/PhysRevD.111.L081302

    doi:10.1103/physrevd.111.l081302 2025

  16. [16]

    2019, MNRAS, 485, 3009, doi: 10.1093/mnras/stz510 Herrera-Mart ´ ın, A., Hendry, M., Gonzalez-Morales, A

    Vives-Arias, H. 2019, MNRAS, 485, 3009, doi: 10.1093/mnras/stz510 Herrera-Mart ´ ın, A., Hendry, M., Gonzalez-Morales, A. X., & Ure˜ na-L´ opez, L. A. 2019, The Astrophysical Journal, 872, 11, doi: 10.3847/1538-4357/aafaf0

    doi:10.1093/mnras/stz510 2019

  17. [17]

    2026, arXiv e-prints, arXiv:2601.16818, doi: 10.48550/arXiv.2601.16818

    Hou, S., Xiang, S., Sming Tsai, Y.-L., et al. 2026, arXiv e-prints, arXiv:2601.16818, doi: 10.48550/arXiv.2601.16818

    doi:10.48550/arxiv.2601.16818 2026

  18. [18]

    2000, PhRvL, 85, 1158, doi: 10.1103/PhysRevLett.85.1158

    Hu, W., Barkana, R., & Gruzinov, A. 2000, PhRvL, 85, 1158, doi: 10.1103/PhysRevLett.85.1158

    doi:10.1103/physrevlett.85.1158 2000

  19. [19]

    2021, ARA&A, 59, 247, doi: 10.1146/annurev-astro-120920-010024

    Hui, L. 2021, ARA&A, 59, 247, doi: 10.1146/annurev-astro-120920-010024

    doi:10.1146/annurev-astro-120920-010024 2021

  20. [21]

    P., Tremaine, S., & Witten, E

    Hui, L., Ostriker, J. P., Tremaine, S., & Witten, E. 2017b, PhRvD, 95, 043541, doi: 10.1103/PhysRevD.95.043541 Irˇ siˇ c, V., Viel, M., Haehnelt, M. G., Bolton, J. S., &

    doi:10.1103/physrevd.95.043541

  21. [22]

    Becker, G. D. 2017, PhRvL, 119, 031302, doi: 10.1103/PhysRevLett.119.031302

    doi:10.1103/physrevlett.119.031302 2017

  22. [23]

    R., Gaudi, B

    Keeton, C. R., Gaudi, B. S., & Petters, A. O. 2003, The Astrophysical Journal, 598, 138–161, doi: 10.1086/378934

    doi:10.1086/378934 2003

  23. [24]

    V., & Khokhlov, A

    Klypin, A., Gottl¨ ober, S., Kravtsov, A. V., & Khokhlov, A. M. 1999a, The Astrophysical Journal, 516, 530, doi: 10.1086/307122

    doi:10.1086/307122

  24. [25]

    V., Valenzuela, O., & Prada, F

    Klypin, A., Kravtsov, A. V., Valenzuela, O., & Prada, F. 1999b, The Astrophysical Journal, 522, 82, doi: 10.1086/307643

    doi:10.1086/307643

  25. [26]

    S., & Dalal, N

    Kochanek, C. S., & Dalal, N. 2004, The Astrophysical Journal, 610, 69–79, doi: 10.1086/421436

    doi:10.1086/421436 2004

  26. [27]

    2022, Monthly Notices of the Royal Astronomical Society, 517, 1867–1883, doi: 10.1093/mnras/stac2677

    Laroche, A., Gilman, D., Li, X., Bovy, J., & Du, X. 2022, Monthly Notices of the Royal Astronomical Society, 517, 1867–1883, doi: 10.1093/mnras/stac2677

    doi:10.1093/mnras/stac2677 2022

  27. [28]

    2025, Phys

    Liao, P.-Y., Su, G.-M., Schive, H.-Y., et al. 2025, Phys. Rev. Lett., 135, 061002, doi: 10.1103/9dqj-q6mt

    doi:10.1103/9dqj-q6mt 2025

  28. [29]

    2024, PhRvD, 110, 083536, doi: 10.1103/PhysRevD.110.083536

    Liu, J., Gao, Z., Biesiada, M., & Liao, K. 2024, PhRvD, 110, 083536, doi: 10.1103/PhysRevD.110.083536

    doi:10.1103/physrevd.110.083536 2024

  29. [30]

    2026, arXiv e-prints, arXiv:2606.06969, doi: 10.48550/arXiv.2606.06969

    Liu, J., Gong, Y., & Zhou, X. 2026, arXiv e-prints, arXiv:2606.06969, doi: 10.48550/arXiv.2606.06969

    Pith/arXiv arXiv doi:10.48550/arxiv.2606.06969 2026

  30. [31]

    P., & Weller, J

    Mao, S., Jing, Y., Ostriker, J. P., & Weller, J. 2004, The Astrophysical Journal, 604, L5–L8, doi: 10.1086/383413

    doi:10.1086/383413 2004

  31. [32]

    1998, MNRAS, 295, 587, doi: 10.1046/j.1365-8711.1998.01319.x

    Mao, S., & Schneider, P. 1998, MNRAS, 295, 587, doi: 10.1046/j.1365-8711.1998.01319.x

    arXiv 1998

  32. [33]

    B., & Madau, P

    Metcalf, R. B., & Madau, P. 2001, ApJ, 563, 9, doi: 10.1086/323695

    doi:10.1086/323695 2001

  33. [34]

    B., & Zhao, H

    Metcalf, R. B., & Zhao, H. 2002, The Astrophysical Journal, 567, L5–L8, doi: 10.1086/339798

    doi:10.1086/339798 2002

  34. [35]

    Miller, John H., J., & Williams, L. L. R. 2025, MNRAS, 543, 3952, doi: 10.1093/mnras/staf1650

    doi:10.1093/mnras/staf1650 2025

  35. [36]

    2017, MNRAS, 471, 4559, doi: 10.1093/mnras/stx1887

    Mocz, P., Vogelsberger, M., Robins, D., et al. 2017, MNRAS, 471, 4559, doi: 10.1093/mnras/stx1887

    doi:10.1093/mnras/stx1887 2017

  36. [37]

    O., Drlica-Wagner, A., Bechtol, K., et al

    Nadler, E. O., Drlica-Wagner, A., Bechtol, K., et al. 2021, Phys. Rev. Lett., 126, 091101, doi: 10.1103/PhysRevLett.126.091101 Planck Collaboration, Aghanim, N., Akrami, Y., et al. 2020, A&A, 641, A6, doi: 10.1051/0004-6361/201833910

    doi:10.1103/physrevlett.126.091101 2021

  37. [38]

    M., Vegetti, S., McKean, J

    Powell, D. M., Vegetti, S., McKean, J. P., et al. 2023, MNRAS, 524, L84, doi: 10.1093/mnrasl/slad074

    doi:10.1093/mnrasl/slad074 2023

  38. [39]

    2024, Phys

    Pozo, A., Broadhurst, T., de Martino, I., et al. 2024, Phys. Rev. D, 110, 043534, doi: 10.1103/PhysRevD.110.043534 18Zhou et al

    doi:10.1103/physrevd.110.043534 2024

  39. [40]

    2002, A&A, 382, L26, doi: 10.1051/0004-6361:20011664

    Reimers, D., Hagen, H.-J., Baade, R., Lopez, S., & Tytler, D. 2002, A&A, 382, L26, doi: 10.1051/0004-6361:20011664

    doi:10.1051/0004-6361:20011664 2002

  40. [41]

    K., & Peiris, H

    Rogers, K. K., & Peiris, H. V. 2021, Phys. Rev. Lett., 126, 071302, doi: 10.1103/PhysRevLett.126.071302

    doi:10.1103/physrevlett.126.071302 2021

  41. [42]

    Safarzadeh, M., & Spergel, D. N. 2020, ApJ, 893, 21, doi: 10.3847/1538-4357/ab7db2

    doi:10.3847/1538-4357/ab7db2 2020

  42. [43]

    2026, Living Reviews in Computational Astrophysics, 12, 1, doi: 10.1007/s41115-026-00027-5

    Schive, H.-Y. 2026, Living Reviews in Computational Astrophysics, 12, 1, doi: 10.1007/s41115-026-00027-5

    doi:10.1007/s41115-026-00027-5 2026

  43. [44]

    2014a, Nature Physics, 10, 496, doi: 10.1038/nphys2996

    Schive, H.-Y., Chiueh, T., & Broadhurst, T. 2014a, Nature Physics, 10, 496, doi: 10.1038/nphys2996

    doi:10.1038/nphys2996

  44. [45]

    2014b, PhRvL, 113, 261302, doi: 10.1103/PhysRevLett.113.261302 Sch¨ onemann, P

    Schive, H.-Y., Liao, M.-H., Woo, T.-P., et al. 2014b, PhRvL, 113, 261302, doi: 10.1103/PhysRevLett.113.261302 Sch¨ onemann, P. H. 1966, Psychometrika, 31, 1, doi: 10.1007/BF02289451

    doi:10.1103/physrevlett.113.261302 1966

  45. [46]

    C., & Engels, J

    Schwabe, B., Niemeyer, J. C., & Engels, J. F. 2016, PhRvD, 94, 043513, doi: 10.1103/PhysRevD.94.043513

    doi:10.1103/physrevd.94.043513 2016

  46. [47]

    J., Birrer, S., Treu, T., et al

    Shajib, A. J., Birrer, S., Treu, T., et al. 2018, Monthly Notices of the Royal Astronomical Society, 483, 5649–5671, doi: 10.1093/mnras/sty3397

    doi:10.1093/mnras/sty3397 2018

  47. [48]

    R., & Ablowitz, M

    Taha, T. R., & Ablowitz, M. I. 1984, Journal of Computational Physics, 55, 203, doi: 10.1016/0021-9991(84)90003-2

    doi:10.1016/0021-9991(84)90003-2 1984

  48. [49]

    Vegetti, S., Czoske, O., Koopmans, L. V. E., & McKean, J. P. 2010, MNRAS, 407, 225, doi: 10.1111/j.1365-2966.2010.16952.x

    arXiv 2010

  49. [50]

    Vegetti, S., & Koopmans, L. V. E. 2009, MNRAS, 392, 945, doi: 10.1111/j.1365-2966.2008.14005.x

    arXiv 2009

  50. [51]

    C., & Schwabe, B

    Veltmaat, J., Niemeyer, J. C., & Schwabe, B. 2018, PhRvD, 98, 043509, doi: 10.1103/PhysRevD.98.043509

    doi:10.1103/physrevd.98.043509 2018

  51. [52]

    2015, Monthly Notices of the Royal Astronomical Society, 447, 3189–3206, doi: 10.1093/mnras/stu2673

    Xu, D., Sluse, D., Gao, L., et al. 2015, Monthly Notices of the Royal Astronomical Society, 447, 3189–3206, doi: 10.1093/mnras/stu2673

    doi:10.1093/mnras/stu2673 2015

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.