Pith. sign in

REVIEW 5 minor 48 references

A comparison of numerical schemes for driven subsonic MHD turbulence

T0 review · 0 major / 5 minor · reviewed 2026-07-10 · grok-4.5

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

Pith's one-line read In driven subsonic MHD turbulence boxes, every tested code and divergence scheme amplifies a seed field exponentially to saturation; residual differences track numerical diffusion, and constrained transport shows no systematic edge over cle

desk verdict Careful multi-code comparison shows CT and cleaning give consistent small-scale dynamos in this idealised subsonic box; residual differences track numerical diffusivity. read the letter →

arxiv 2607.06660 v1 pith:XH6APPNB submitted 2026-07-07 astro-ph.IM

Rüdiger Pakmor , Thomas Guillet , Rebekka Bieri , Christoph Pfrommer , Volker Springel , Romain Teyssier This is my paper
classification astro-ph.IM
keywords MHDsmall-scaledynamonumericalmethodsconstrainedtransportdivergencecleaningdriventurbulencesubsoniccodecomparison
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Astrophysical simulations rarely resolve the full range of scales on which a small-scale dynamo amplifies magnetic fields, so the growth of seed fields is controlled by numerical dissipation. This paper isolates that effect in the simplest useful setting: identical periodic boxes of driven subsonic turbulence with a weak seed field and purely numerical viscosity and resistivity. Three widely used MHD codes are run with the same driving and the same analysis pipeline, covering constrained-transport, Powell, and Dedner divergence treatments. At sufficient resolution every scheme produces the expected exponential growth until saturation, and the structural properties of the magnetic field match theoretical expectations and one another. Differences that remain in growth rate and saturation level at fixed resolution are explained by how much numerical diffusion each scheme injects. For this idealised problem the authors find no systematic advantage of formally divergence-free methods over divergence-cleaning methods.

What carries the argument

Identical solenoidal turbulent driving (same Ornstein–Uhlenbeck realisation, same power spectrum, same force density) together with a shared analysis pipeline applied to Arepo, Athena, and Ramses runs that differ only in hydrodynamics and divergence-control scheme (constrained transport, Powell cleaning, Dedner cleaning).

What would settle it

A controlled comparison of the same codes with explicit viscosity and resistivity at matched Reynolds and Prandtl numbers, or the same codes run on driven supersonic turbulence or a multi-phase ISM box: if CT then systematically outperforms cleaning (or vice versa) under those more realistic conditions, the present claim of no systematic advantage would be falsified for the regimes that actually matter.

Watch

Extended reading notes

Core claim

At sufficient resolution every code and scheme tested exhibits dynamo-like exponential amplification of a weak seed field until saturation; residual differences in kinematic amplification rates and saturation strengths are attributable to varying levels of numerical diffusion, and for this idealised driven-subsonic setup there is no systematic advantage of constrained transport over divergence-cleaning methods.

Load-bearing premise

The claim that pure numerical dissipation on an idealised, isothermal, subsonic, periodic box is a good enough proxy for the numerical regime of multi-scale astrophysical simulations that conclusions about constrained transport versus cleaning can be transferred.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 5 minor

Summary. The paper compares six MHD implementations (Arepo-Powell, Arepo-Dedner, Athena-CT, Ramses-CT, DG-Powell, DG-Dedner) on identical driven subsonic isothermal turbulence boxes with a weak seed field and purely numerical dissipation, at resolutions 32^{3}–256^{3}. Using the same solenoidal Ornstein–Uhlenbeck driving and a common analysis pipeline, the authors measure kinematic growth rates Γ, saturation Emag/Ekin, structural wavenumbers (k∥, kB imes J, krms, kλ, kB·J), histograms of B, J^{2} and curvature K, B–K correlations, and kinetic/magnetic power spectra in both kinematic and saturated regimes. At sufficient resolution every scheme produces exponential dynamo growth to saturation; residual differences in rates and saturation levels track numerical diffusivity (illustrated by matched spectra, e.g. Arepo-Powell 128^{3} vs Ramses-CT 256^{3} at Emag/Ekin = 10^{-2}); structural diagnostics do not systematically separate constrained-transport from divergence-cleaning methods. Arepo-CT is shown to be unstable at high resolution and is excluded. The authors repeatedly flag the idealised setup as a first step.

Significance. If the result holds within its stated scope, the paper supplies a carefully controlled, multi-code benchmark that is directly useful to the astrophysical MHD community. The identical driving realisation, shared analysis pipeline, multi-resolution suite, and suite of diagnostics (growth rates, saturation, structure functions, power spectra, B–J–K histograms) make the comparison unusually clean. Explicit documentation of the Arepo-CT instability and the quantitative matching of spectra across codes at different resolutions are concrete, reusable contributions. The finding of no systematic CT advantage for this subsonic, numerically dissipative setup is a useful data point for code choice in galaxy-formation and ISM contexts, provided the idealisation is kept in view.

minor comments (5)
  1. Table 1: the column header for the growth-rate interval uses t0 and t1 without units; a brief reminder that times are in code units (sound-crossing time of the box) would help readers who jump straight to the table.
  2. §2.1 / §6.2: the Dedner parameters (ch, τ or cr) and the Athena slope-limiter choice (characteristic vs primitive) are free parameters that affect resistivity; a short sentence noting that the chosen values are the codes’ standard defaults would make the comparison more transparent.
  3. Figure 5 caption and §4: the statement that DG-Powell resolves the smallest structures is clear from the slices, but a quantitative measure (e.g. the k at which the magnetic spectrum rolls off) would make the visual impression more precise.
  4. §6.3: the discussion of Arepo-CT instabilities is valuable; a one-sentence pointer to the forthcoming Springel et al. (2026) formulation would help readers who need a stable moving-mesh CT option.
  5. References: Tomida et al. (2026) is cited for divergence-control tests; ensuring the arXiv identifier is complete and consistent with the rest of the bibliography would avoid lookup friction.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: empirical multi-code comparison with directly measured growth rates, spectra and structural diagnostics.

full rationale

The paper is a controlled numerical experiment comparing MHD codes and divergence schemes on identical driven subsonic boxes. Amplification rates Γ are measured from the simulation time series (times when Emag/Ekin first crosses 10^{-4} and 10^{-2}), saturation levels are time averages of Emag/Ekin, and structural diagnostics (k∥, kB imes J, histograms of B/J^{2}/K, power spectra) are computed from the same snapshots. These quantities are not derived from fitted parameters that reappear as predictions, nor are they forced by self-definition. Self-citations (Pakmor et al., Guillet et al., Mocz et al., etc.) supply code implementations and prior method papers; they do not underwrite the dynamo results themselves. Theoretical expectations (Kazantsev k^{3/2}, B∝K^{-1/2}) are external literature, not author uniqueness theorems. The idealised setup is repeatedly flagged as a limitation rather than smuggled into a broader claim. Within the stated scope the derivation chain is self-contained and non-circular.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central claim rests on standard MHD equations, the numerical implementations of three production codes, and a handful of conventional choices for driving and thermodynamics. No new physical entities are postulated; free parameters are the usual numerical knobs (Courant numbers, cleaning speeds, driving amplitude) that are fixed once and for all rather than fitted to the dynamo outcome.

free parameters (4)
  • driving amplitude √P0 = 2e-4
    Set by hand to 2×10^{-4} so that the steady-state Mach number is ≈0.2; not fitted to magnetic-field data.
  • Dedner cleaning parameters (c_h, τ or c_r)
    Chosen once per code following original Dedner prescriptions or simple generalisations; not optimised against the dynamo growth rate.
  • Courant factors and slope-limiter choices
    Standard production values for each code; varied only in a controlled Athena sensitivity test.
  • adiabatic index γ=1.01 and isothermal reset = 1.01
    Imposed to keep the flow nearly isothermal; conventional for this class of test.
assumptions (3)
  • domain assumption Ideal MHD equations with purely numerical viscosity and resistivity adequately capture the small-scale dynamo in the low-resolution regime of galaxy-formation simulations.
    Stated in the introduction and methods; the entire comparison is performed under this premise.
  • domain assumption Identical solenoidal Ornstein–Uhlenbeck driving and identical post-processing remove code-to-code bias sufficiently that residual differences can be attributed to numerical diffusion.
    Core methodological claim of §2.2–2.3.
  • standard math Standard second-order finite-volume / DG / CT discretisations of the induction equation are consistent with the continuous MHD equations in the continuum limit.
    Background numerical-analysis result used throughout.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A comparison of numerical schemes for driven subsonic MHD turbulence." pith.science (2026). https://pith.science/paper/XH6APPNB

@misc{pith2026260706660,
  author       = {Pith},
  title        = {Pith review of: A comparison of numerical schemes for driven subsonic MHD turbulence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XH6APPNB}},
  note         = {Machine review of arXiv:2607.06660}
}
read the original abstract

Turbulence is ubiquitous in astrophysical systems, and since most cosmic gas is ionised, it supports magnetic fields. In turbulent environments, these fields are rapidly amplified through a small-scale dynamo. Multi-scale astrophysical simulations, however, rarely resolve this process adequately. Limited spatial dynamic range makes small-scale amplification sensitive to the numerical choices made in the hydrodynamics and magnetic field solvers. Here, we investigate idealised periodic boxes of driven subsonic turbulence with a weak seed magnetic field. These simulations with purely numerical dissipation provide a simple environment in which an efficient small-scale dynamo is expected. We aim to systematically compare the three widely used magnetohydrodynamics (MHD) codes \textsc{Arepo}, \textsc{Athena}, and \textsc{Ramses} across different divergence-control schemes: constrained transport, Powell cleaning, and Dedner cleaning. To minimise comparison bias, we adopt identical turbulent driving and analysis pipelines across all runs. At sufficient resolution, every code and scheme we test exhibits dynamo-like exponential amplification of the seed field until saturation. The structural properties of the magnetic field in both the kinematic and saturated regimes are consistent across schemes (with the exception of \textsc{Arepo}'s constrained transport) and agree with theoretical expectations. Residual differences, particularly in kinematic amplification rates and saturation strengths at fixed resolution, appear attributable to varying levels of numerical diffusion. Notably, we find for this setup no systematic advantage of constrained transport over divergence-cleaning methods. We stress that this comparison, conducted in a highly idealised setting, represents a first step. Future extensions to more complex and physically realistic configurations remain essential.

Figures

Figures reproduced from arXiv: 2607.06660 by the authors.

Figure 1
Figure 1. Amplification of magnetic fields over time. The panels show the time evolution of the ratio of magnetic to kinetic energy for all four resolution levels from 323 (top left) to 2563 (bottom right). Each panel shows one line for each code. The gray horizontal lines show the threshold values between which the magnetic field amplification rate is fitted in the kinematic regime. The gray shaded area on the right of each … view at source ↗
Figure 2
Figure 2. The left panel shows the time evolution of the kinetic energy in the box focusing on the early time of the simulation 𝑡 < 20 for all codes at a resolution of 323 and 2563 . All simulations have saturated the kinetic energy at 𝑡 = 5, and it remains roughly constant afterwards. The right panel shows the mean Mach number M for different resolutions and numerical schemes; see the text for the definition for M. The value… view at source ↗
Figure 3
Figure 3. Exponential amplification rate of the magnetic energy in the kine￾matic regime for different resolutions and numerical schemes. The rate is measured between the times when the ratio of magnetic energy to kinetic en￾ergy in the box first reaches 10−4 and 10−2 (see horizontal lines in [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (10 more)
Figure 5
Figure 5. Figure 5: Slices of magnetic field strength at 𝑡 = 100 for all six schemes at the highest numerical resolution of 2563 . The overall field strength and large scale structures are similar, but the amount of small scale structure varies significantly. DG-Powell shows the smallest …
Figure 6
Figure 6. Figure 6: Various global measures of structural properties of the magnetic field in the saturated regime for different resolutions and numerical schemes. We show time averages of the quantities between 𝑡 = 80 and 𝑡 = 100. The different schemes span a range of values at fixed num…
Figure 7
Figure 7. Figure 7: Time averaged volume-weighted histograms of the log10 of the magnetic field strength 𝐵, current squared 𝐽 2 , and curvature 𝐾 of the saturated magnetic field for the different numerical schemes at a numerical resolution of 2563 . We stack histograms between 𝑡 = 80 and …
Figure 8
Figure 8. Figure 8: Contours of the time averaged volume weighted 2D histogram of magnetic field strength 𝐵 versus curvature 𝐾 in the saturated regime. The histograms are for the different numerical schemes at a numerical resolution of 2563 and stacked between 𝑡 = 80 and 𝑡 = 100. We inclu…
Figure 10
Figure 10. Figure 10: Magnetic power spectra for different schemes in the kinematic regime at a resolution of 2563 . We show two simulation times for each scheme, when the total magnetic energy first surpasses 10−4 and 10−2 of the total kinetic energy, respectively. The shaded grey area in…
Figure 11
Figure 11. Figure 11: Magnetic power spectra for Ramses-CT and Arepo-Powell in the kinematic regime at the time when the total magnetic energy first surpasses 10−2 of the total kinetic energy for different numerical resolutions. The shaded grey area indicates the scales on which we drive t…
Figure 12
Figure 12. Figure 12: Magnetic and kinetic power spectra for Arepo-Powell in the saturated state for different numerical resolutions. The shaded grey area indicates the scales on which we drive turbulence. On large scales the power spectra are nearly converged even for the lowest resolutio…
Figure 13
Figure 13. Figure 13: Comparison of Athena with different settings for reconstruction and slope limiter, including a second order scheme with linear reconstruction (LR), a higher order version using third order time integration and the piecewise parabolic method for reconstruction (PPM), a…
Figure 14
Figure 14. Figure 14: Slices of the magnetic field strength at 𝑡 = 100 (left column) and density (right column) for Arepo-Powell (at a resolution of 2563 ) and Arepo-CT at resolutions of 1283 and 2563 . At 2563 the Arepo-CT scheme becomes obviously unstable and shows prominent oscillatory …
Figure 15
Figure 15. Figure 15: Histograms of the log10 of the magnetic field strength 𝐵, velocity magnitude 𝑉, and density 𝜌 for Arepo-CT at different numerical resolutions, similar to [PITH_FULL_IMAGE:figures/full_fig_p015_15.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages

  1. [1]

    S., 2012, @doi [Journal of Computational Physics] 10.1016/j.jcp.2011.12.025 , https://ui.adsabs.harvard.edu/abs/2012JCoPh.231.7476B 231, 7476

    Balsara D. S., 2012, @doi [Journal of Computational Physics] 10.1016/j.jcp.2011.12.025 , https://ui.adsabs.harvard.edu/abs/2012JCoPh.231.7476B 231, 7476

    work page doi:10.1016/j.jcp.2011.12.025 2012

  2. [2]

    Bauer A., Springel V., 2012, @doi [Monthly Notices of the Royal Astronomical Society] 10.1111/j.1365-2966.2012.21058.x , 423, 2558

    work page doi:10.1111/j.1365-2966.2012.21058.x 2012

  3. [3]

    R., Federrath C., Kriel N., Mocz P., Seta A., 2023, @doi [ ] 10.1093/mnras/stad1863 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.524.3201B 524, 3201

    Beattie J. R., Federrath C., Kriel N., Mocz P., Seta A., 2023, @doi [ ] 10.1093/mnras/stad1863 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.524.3201B 524, 3201

    work page doi:10.1093/mnras/stad1863 2023

  4. [4]

    Brandenburg A., 2011, @doi [The Astrophysical Journal] 10.1088/0004-637X/741/2/92 , 741, 92

    work page doi:10.1088/0004-637x/741/2/92 2011

  5. [5]

    Brandenburg A., Subramanian K., 2005, @doi [ ] 10.1016/j.physrep.2005.06.005 , https://ui.adsabs.harvard.edu/abs/2005PhR...417....1B 417, 1

    work page doi:10.1016/j.physrep.2005.06.005 2005

  6. [6]

    Brandenburg A., Sokoloff D., Subramanian K., 2012, @doi [ ] 10.1007/s11214-012-9909-x , https://ui.adsabs.harvard.edu/abs/2012SSRv..169..123B 169, 123

    work page doi:10.1007/s11214-012-9909-x 2012

  7. [7]

    Cernetic M., Springel V., Guillet T., Pakmor R., 2023, @doi [ ] 10.1093/mnras/stad1043 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.522..982C 522, 982

    work page doi:10.1093/mnras/stad1043 2023

  8. [8]

    Cernetic M., Springel V., Guillet T., Pakmor R., 2024, @doi [ ] 10.1093/mnras/stae2192 , https://ui.adsabs.harvard.edu/abs/2024MNRAS.534.1963C 534, 1963

    work page doi:10.1093/mnras/stae2192 2024

Show all 48 references
  1. [9]

    R., 1984, @doi [Journal of Computational Physics] 10.1016/0021-9991(84)90143-8 , https://ui.adsabs.harvard.edu/abs/1984JCoPh..54..174C 54, 174

    Colella P., Woodward P. R., 1984, @doi [Journal of Computational Physics] 10.1016/0021-9991(84)90143-8 , https://ui.adsabs.harvard.edu/abs/1984JCoPh..54..174C 54, 174

    1984 doi

  2. [10]

    A., van de Voort F., 2023, @doi [ ] 10.1146/annurev-astro-041923-043618 , https://ui.adsabs.harvard.edu/abs/2023ARA&A..61..473C 61, 473

    Crain R. A., van de Voort F., 2023, @doi [ ] 10.1146/annurev-astro-041923-043618 , https://ui.adsabs.harvard.edu/abs/2023ARA&A..61..473C 61, 473

    2023 doi

  3. [11]

    Dedner A., Kemm F., Kr \"o ner D., Munz C.-D., Schnitzer T., Wesenberg M., 2002, @doi [Journal of Computational Physics] 10.1006/jcph.2001.6961 , https://ui.adsabs.harvard.edu/abs/2002JCoPh.175..645D 175, 645

    2002 doi

  4. [12]

    G., Scalo J., 2004, @doi [ ] 10.1146/annurev.astro.41.011802.094859 , https://ui.adsabs.harvard.edu/abs/2004ARA&A..42..211E 42, 211

    Elmegreen B. G., Scalo J., 2004, @doi [ ] 10.1146/annurev.astro.41.011802.094859 , https://ui.adsabs.harvard.edu/abs/2004ARA&A..42..211E 42, 211

    2004 doi arXiv

  5. [13]

    R., Hawley J

    Evans C. R., Hawley J. F., 1988, @doi [ ] 10.1086/166684 , https://ui.adsabs.harvard.edu/abs/1988ApJ...332..659E 332, 659

    1988 doi

  6. [14]

    S., Schmidt W., Mac Low M.-M., 2010, @doi [Astronomy and Astrophysics] 10.1051/0004-6361/200912437 , 512, A81

    Federrath C., Roman-Duval J., Klessen R. S., Schmidt W., Mac Low M.-M., 2010, @doi [Astronomy and Astrophysics] 10.1051/0004-6361/200912437 , 512, A81

    2010 doi

  7. [15]

    pp 576--599 ( @eprint arXiv 2507.08925 ), @doi 10.1016/B978-0-443-21439-4.00111-5

    Feldmann R., Bieri R., 2026, in Encyclopedia of Astrophysics, Volume 4. pp 576--599 ( @eprint arXiv 2507.08925 ), @doi 10.1016/B978-0-443-21439-4.00111-5

    2026 doi arXiv

  8. [16]

    Gottlieb S., 2005, @doi [Journal of Scientific Computing] 10.1007/s10915-004-4635-5 , 25, 105

    2005 doi

  9. [17]

    W., Beckwith K., 2023, @doi [ ] 10.3847/2041-8213/acaea7 , https://ui.adsabs.harvard.edu/abs/2023ApJ...942L..34G 942, L34

    Grete P., O'Shea B. W., Beckwith K., 2023, @doi [ ] 10.3847/2041-8213/acaea7 , https://ui.adsabs.harvard.edu/abs/2023ApJ...942L..34G 942, L34

    2023 doi

  10. [18]

    Guillet T., Pakmor R., Springel V., Chandrashekar P., Klingenberg C., 2019, @doi [ ] 10.1093/mnras/stz314 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.485.4209G 485, 4209

    2019 doi

  11. [19]

    P., Ruzmaikin A

    Kazantsev A. P., Ruzmaikin A. A., Sokolov D. D., 1985, Zhurnal Eksperimentalnoi i Teoreticheskoi Fiziki, https://ui.adsabs.harvard.edu/abs/1985ZhETF..88..487K 88, 487

    1985

  12. [20]

    R., Seta A., Federrath C., 2022, @doi [ ] 10.1093/mnras/stac969 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.513.2457K 513, 2457

    Kriel N., Beattie J. R., Seta A., Federrath C., 2022, @doi [ ] 10.1093/mnras/stac969 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.513.2457K 513, 2457

    2022 doi

  13. [21]

    G., et al., 2011, @doi [The Astrophysical Journal] 10.1088/0004-637X/737/1/13 , 737, 13

    Kritsuk A. G., et al., 2011, @doi [The Astrophysical Journal] 10.1088/0004-637X/737/1/13 , 737, 13

    2011 doi

  14. [22]

    Martin-Alvarez S., Devriendt J., Slyz A., Teyssier R., 2018, @doi [ ] 10.1093/mnras/sty1623 , https://ui.adsabs.harvard.edu/abs/2018MNRAS.479.3343M 479, 3343

    2018 doi

  15. [23]

    Miyoshi T., Kusano K., 2005, @doi [Journal of Computational Physics] 10.1016/j.jcp.2005.02.017 , https://ui.adsabs.harvard.edu/abs/2005JCoPh.208..315M 208, 315

    2005 doi

  16. [24]

    Mocz P., Pakmor R., Springel V., Vogelsberger M., Marinacci F., Hernquist L., 2016, @doi [ ] 10.1093/mnras/stw2004 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.463..477M 463, 477

    2016 doi

  17. [25]

    Pakmor R., Springel V., 2013, @doi [ ] 10.1093/mnras/stt428 , https://ui.adsabs.harvard.edu/abs/2013MNRAS.432..176P 432, 176

    2013 doi

  18. [26]

    Pakmor R., Bauer A., Springel V., 2011, @doi [ ] 10.1111/j.1365-2966.2011.19591.x , https://ui.adsabs.harvard.edu/abs/2011MNRAS.418.1392P 418, 1392

    2011 doi

  19. [27]

    J., Ohlmann S

    Pakmor R., Springel V., Bauer A., Mocz P., Munoz D. J., Ohlmann S. T., Schaal K., Zhu C., 2016, @doi [ ] 10.1093/mnras/stv2380 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.455.1134P 455, 1134

    2016 doi

  20. [28]

    Pakmor R., et al., 2017, @doi [ ] 10.1093/mnras/stx1074 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.469.3185P 469, 3185

    2017 doi

  21. [29]

    Pakmor R., et al., 2024, @doi [ ] 10.1093/mnras/stae112 , https://ui.adsabs.harvard.edu/abs/2024MNRAS.528.2308P 528, 2308

    2024 doi

  22. [30]

    M., 2022, @doi [ ] 10.1093/mnras/stac1808 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.515.4229P 515, 4229

    Pfrommer C., Werhahn M., Pakmor R., Girichidis P., Simpson C. M., 2022, @doi [ ] 10.1093/mnras/stac1808 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.515.4229P 515, 4229

    2022 doi

  23. [31]

    G., Roe P

    Powell K. G., Roe P. L., Linde T. J., Gombosi T. I., De Zeeuw D. L., 1999, @doi [Journal of Computational Physics] 10.1006/jcph.1999.6299 , https://ui.adsabs.harvard.edu/abs/1999JCoPh.154..284P 154, 284

    1999 doi

  24. [32]

    Rieder M., Teyssier R., 2016, @doi [ ] 10.1093/mnras/stv2985 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.457.1722R 457, 1722

    2016 doi

  25. [33]

    Rieder M., Teyssier R., 2017a, @doi [ ] 10.1093/mnras/stx1670 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.471.2674R 471, 2674

    doi

  26. [34]

    Rieder M., Teyssier R., 2017b, @doi [ ] 10.1093/mnras/stx2276 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.472.4368R 472, 4368

    doi

  27. [35]

    Schaal K., Bauer A., Chandrashekar P., Pakmor R., Klingenberg C., Springel V., 2015, @doi [ ] 10.1093/mnras/stv1859 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.453.4278S 453, 4278

    2015 doi

  28. [36]

    A., Cowley S

    Schekochihin A. A., Cowley S. C., Taylor S. F., Maron J. L., McWilliams J. C., 2004, @doi [ ] 10.1086/422547 , https://ui.adsabs.harvard.edu/abs/2004ApJ...612..276S 612, 276

    2004 doi

  29. [37]

    C., 2006, @doi [Computers & Fluids] 10.1016/j.compfluid.2005.03.002 , 35, 353

    Schmidt W., Hillebrandt W., Niemeyer J. C., 2006, @doi [Computers & Fluids] 10.1016/j.compfluid.2005.03.002 , 35, 353

    2006 doi

  30. [38]

    Springel V., 2010, @doi [ ] 10.1111/j.1365-2966.2009.15715.x , https://ui.adsabs.harvard.edu/abs/2010MNRAS.401..791S 401, 791

    2010 doi

  31. [39]

    M., Gardiner T

    Stone J. M., Gardiner T. A., Teuben P., Hawley J. F., Simon J. B., 2008, @doi [ ] 10.1086/588755 , https://ui.adsabs.harvard.edu/abs/2008ApJS..178..137S 178, 137

    2008 doi

  32. [40]

    M., Tomida K., White C

    Stone J. M., Tomida K., White C. J., Felker K. G., 2020, @doi [ ] 10.3847/1538-4365/ab929b , https://ui.adsabs.harvard.edu/abs/2020ApJS..249....4S 249, 4

    2020 doi

  33. [41]

    Teyssier R., 2002, @doi [ ] 10.1051/0004-6361:20011817 , https://ui.adsabs.harvard.edu/abs/2002A&A...385..337T 385, 337

    2002 doi

  34. [42]

    Teyssier R., Fromang S., Dormy E., 2006, @doi [Journal of Computational Physics] 10.1016/j.jcp.2006.01.042 , https://ui.adsabs.harvard.edu/abs/2006JCoPh.218...44T 218, 44

    2006 doi

  35. [43]

    E., Takasao S., Iwasaki K., 2026, @doi [arXiv e-prints] 10.48550/arXiv.2605.07928 , https://ui.adsabs.harvard.edu/abs/2026arXiv260507928T p

    Tomida K., Sadanari K. E., Takasao S., Iwasaki K., 2026, @doi [arXiv e-prints] 10.48550/arXiv.2605.07928 , https://ui.adsabs.harvard.edu/abs/2026arXiv260507928T p. arXiv:2605.07928

    2026 doi arXiv

  36. [44]

    S., Werk J

    Tumlinson J., Peeples M. S., Werk J. K., 2017, @doi [ ] 10.1146/annurev-astro-091916-055240 , https://ui.adsabs.harvard.edu/abs/2017ARA&A..55..389T 55, 389

    2017 doi

  37. [45]

    Vogelsberger M., Sijacki D., Kere s D., Springel V., Hernquist L., 2012, @doi [ ] 10.1111/j.1365-2966.2012.21590.x , https://ui.adsabs.harvard.edu/abs/2012MNRAS.425.3024V 425, 3024

    2012 doi

  38. [46]

    Vogelsberger M., Marinacci F., Torrey P., Puchwein E., 2020, @doi [Nature Reviews Physics] 10.1038/s42254-019-0127-2 , https://ui.adsabs.harvard.edu/abs/2020NatRP...2...42V 2, 42

    2020 doi

  39. [47]

    Weinberger R., Springel V., Pakmor R., 2020, @doi [ ] 10.3847/1538-4365/ab908c , https://ui.adsabs.harvard.edu/abs/2020ApJS..248...32W 248, 32

    2020 doi

  40. [48]

    Wissing R., Shen S., 2023, @doi [ ] 10.1051/0004-6361/202244753 , https://ui.adsabs.harvard.edu/abs/2023A&A...673A..47W 673, A47

    2023 doi

Pith tools

Reviewed July 10, 2026 · model on record in the stance chip above.