Pith. sign in

REVIEW 3 major objections 7 minor 42 references

3D turbulence recovered from noisy 2D maps at 5% accuracy

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 ↗

T0 review · glm-5.2

2026-07-09 17:17 UTC pith:4IWKAZVC

load-bearing objection Useful denoising correction for the Brunt 3D density dispersion method, but the empirical calibration is fit to subsonic/transonic hydrodynamic simulations and its generalization to supersonic magnetized ISM conditions is untested. the 3 major comments →

arxiv 2607.07205 v1 pith:4IWKAZVC submitted 2026-07-08 astro-ph.GA astro-ph.IMastro-ph.SRphysics.comp-phphysics.data-an

From 2D to 3D: Recovering Turbulent Density Dispersions from Noisy Data

Luz L. Jimenez Vela , Christoph Federrath , David C. Collins , Seth Davidovits This is my paper
classification astro-ph.GA astro-ph.IMastro-ph.SRphysics.comp-phphysics.data-an
keywords turbulencedensity dispersionpower spectrumsignal-to-noise ratioBrunt methoddenoisingcolumn densityinterstellar medium
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 Brunt method reconstructs the 3D density dispersion of a turbulent medium from 2D column-density observations, but it breaks down when the data are noisy: at signal-to-noise ratios near unity, the reconstructed dispersion can be overestimated by factors of several. This paper identifies a characteristic wavenumber, k_noise, marking the spatial scale at which the noise power spectrum overtakes the signal power spectrum. By truncating the Brunt reconstruction at k_noise—summing only over wavenumbers where signal dominates—the authors recover the noise-free 3D density dispersion to within 5% for SNR ≥ 3 and within 15% for SNR ≥ 1. The cutoff k_noise is determined empirically from the measurement SNR and image resolution via a two-equation model (Eqs. 13–14), requiring no knowledge of the noise spectrum itself. An alternative correction—direct subtraction of a measured noise spectrum—works even better when the noise is characterized. The method is tested on hydrodynamic simulations with three density-perturbation amplitudes and five noise spectral slopes, and the correction proves robust across all these variations.

Core claim

The signal and noise in a column-density power spectrum intersect at a well-defined wavenumber k_noise that depends primarily on SNR and image resolution, and truncating the Brunt 3D density-dispersion reconstruction at this cutoff removes the noise bias to within 5–15% accuracy down to SNR ≈ 1. The universality of k_noise across different turbulence amplitudes and noise types means the correction can be applied using only the measurement SNR and pixel count, without characterizing the noise spectrum.

What carries the argument

The central mechanism is the spectral intersection: the turbulent density signal has a decreasing power spectrum (more power on large scales), while detector noise contributes a flat or rising spectrum (more power on small scales). These two curves cross at k_noise. Below k_noise, the observed spectrum is signal-dominated; above it, noise-dominated. The Brunt method's reconstruction formula (Eq. 11) sums 2k·P_2D(N,k) over all wavenumbers, so noise power at high k inflates the sum. Restricting the sum to k ≤ k_noise (Eq. 15) excises the noise-dominated tail. The empirical model for k_noise (Eq. 13) is a sigmoid-like function of SNR/SNR_crit, where SNR_crit scales as k_max^(2/3) (Eq. 14), with

Load-bearing premise

The empirical model for the noise cutoff wavenumber k_noise is calibrated on simulations with Mach numbers 0.29–0.64 and is assumed to be largely insensitive to Mach number, turbulence driving mode, and magnetic field strength. However, the density power spectrum slope is known to change with these parameters, which would shift the signal–noise intersection and potentially invalidate the universal k_noise(SNR, k_max) relation at higher Mach numbers or in magnetized turbulence

What would settle it

If the density power spectrum slope varies significantly with Mach number, driving mode, or magnetic field strength—as prior work suggests it does—the signal–noise intersection k_noise would shift, and the empirical relation k_noise(SNR, k_max) calibrated on low-Mach-number simulations would systematically misidentify the cutoff, producing biased density-dispersion estimates for the supersonic, magnetized turbulence typical of real molecular clouds.

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

If this is right

  • Existing column-density maps of molecular clouds with marginal SNR can be reanalyzed to extract 3D density dispersions that were previously considered unreliable, potentially revising estimates of the turbulence driving parameter b and its connection to star formation rates.
  • The method extends the usable SNR range of the Brunt technique from effectively ≳10 down to ≈1, broadening the pool of observational data suitable for turbulence characterization.
  • The k_noise framework could be adapted for other spectral reconstruction problems where a signal with a decreasing power spectrum is contaminated by noise with a flat or rising spectrum, such as velocity field reconstructions from spectral-line data.
  • If the empirical k_noise(SNR, k_max) relation holds at higher Mach numbers, it would provide a universal noise-correction prescription for the majority of molecular-cloud observations; if it does not, the signal-spectrum slope becomes a necessary additional input.

Where Pith is reading between the lines

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

  • The k_noise cutoff acts as an implicit spatial filter whose physical scale depends on SNR: at SNR = 1, only the largest ~10–20% of resolved scales contribute to the reconstruction, meaning the method trades spatial resolution for noise robustness. This trade-off is not a failure but a quantifiable cost of working with noisy data.
  • The empirical scaling SNR_crit ∝ k_max^(2/3) suggests a geometric relationship between the noise power integrated over 2D shells and the signal power at the intersection, but the paper does not derive this scaling from first principles. A theoretical derivation might reveal whether the 2/3 exponent is universal or specific to the density-spectrum slopes in the simulations used.
  • For time-resolved observations of evolving turbulent regions (e.g., protostellar cores), the method could track 3D density dispersion changes over time from 2D data, provided each epoch's SNR is characterized—though the paper does not discuss temporal applications.

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 / 7 minor

Summary. This paper extends the Brunt et al. (2010a) method for reconstructing the 3D turbulent density dispersion (sigma_rho/rho_0) from 2D column-density maps to handle finite signal-to-noise ratio (SNR). The authors identify a characteristic noise wavenumber, k_noise, at the intersection of the signal and noise power spectra, and propose restricting the Brunt reconstruction to k <= k_noise (Eq. 15). An empirical model for k_noise(SNR, k_max) is provided (Eqs. 13-14), and an alternative noise-subtraction method is also presented. The method is tested on hydrodynamic simulations of shock-driven turbulence with three void sizes (Mach numbers 0.29-0.64), five analysis regions, three projections, five noise spectral exponents, and five resolutions. The claimed accuracy is <5% for SNR>=3 and <15% for SNR>=1.

Significance. The problem addressed is of genuine practical importance: the Brunt method is widely used to estimate 3D density variances from 2D column-density observations, and the effect of finite SNR on this method has not been systematically treated. The noise-correction prescription is concrete and immediately applicable. The paper provides falsifiable, quantitative error bars and tests the method across multiple noise types and resolutions (Appendix C, Fig. 13). The alternative noise-spectrum subtraction approach (Section 5.1) is a useful, physically transparent complement. The work is a natural and useful extension of the Brunt framework.

major comments (3)
  1. The central practical prescription rests on the empirical calibration of k_noise via Eqs. (13)-(14), which are fit to simulations with Mach numbers 0.29, 0.43, and 0.64 (all subsonic to transonic, purely hydrodynamic, from a single simulation family). The authors claim a 'certain level of universality' (Section 5.3) and state that k_noise is 'largely insensitive' to the amplitude of density perturbations (Section 6). However, k_noise is defined as the intersection of the signal spectrum P_signal(k) with the noise spectrum P_noise(k). The signal spectrum slope is known to depend on Mach number, turbulence driving mode, and magnetic field strength (Kim & Ryu 2005; Federrath & Klessen 2013, as cited by the authors themselves). A change in the signal slope would shift k_noise at fixed SNR, potentially invalidating the universal k_noise(SNR, k_max) relation. Typical molecular cloud Mach数为 5-
  2. The functional forms of Eq. (13) (with exponent -8/3) and Eq. (14) (SNR_crit proportional to k_max^{2/3}) are empirical fits with no derivation from the physics of the spectral intersection. The four fitted quantities (SNR_crit, the -8/3 exponent, the 1/4 outer exponent, and the 2/3 power-law index) are calibrated on a narrow parameter range. The paper does not provide a physical argument for why these specific functional forms should hold outside the tested regime. This makes it difficult for a reader to assess the domain of validity. Can the authors provide any analytical or semi-analytical motivation for the -8/3 and 2/3 exponents, or at least discuss what physical scaling one would expect from the intersection of a turbulent signal spectrum with a noise spectrum?
  3. The error claims of <5% for SNR>=3 and <15% for SNR>=1 (Abstract, Section 5.4, Fig. 9) are evaluated on the same simulation suite used to calibrate Eqs. (13)-(14). This is a standard empirical calibration, not a circular derivation, but the 'prediction' of sigma_rho/rho_0 is to some extent a fit to the data used to derive the model. The paper would be substantially strengthened by a validation test on an independent dataset or simulation with different physical parameters (e.g., higher Mach number, different driving, or magnetized turbulence). Without such a test, the error bars should be understood as in-sample fitting residuals rather than predictive accuracy on new data. The authors should clarify this distinction explicitly.
minor comments (7)
  1. Section 4.1: The SNR is defined as the ratio of the intrinsic standard deviation of the column density to the standard deviation of the noise field. It would be helpful to clarify whether this is a global SNR (single number for the whole map) or a pixel-by-pixel SNR, and how the definition relates to observationally common definitions (e.g., peak-to-noise or mean-to-noise).
  2. Figure 5: The y-axis label 'P' is ambiguous. It would be helpful to specify whether this is the angle-integrated power spectrum (2*pi*k*P_2D) or the azimuthally averaged P_2D, and to include units or normalization.
  3. Table 2: The column headers do not explicitly state that columns 2-5 are medians with 16th-84th percentile ranges. This is noted in the caption but easy to miss; consider adding a note in the header row.
  4. Section 5.3, last paragraph: The phrase 'a certain level of universality' is vague. Consider quantifying the variation more precisely, or stating the range of parameters over which universality has been tested.
  5. The paper mentions that extensions to non-cubic domains have been explored by Yoon & Cho (2024) and that results can be extended accordingly (Section 2.2), but does not test this. A brief comment on whether the noise correction is expected to be affected by non-cubic geometry would be useful.
  6. Equation (12): The notation uses k for the wavenumber magnitude, but the noise generation uses k^{beta/2}. It would help to clarify that this is the Fourier-space radial coordinate, consistent with the k used in the power spectra.
  7. The abstract states errors of '<~5%' and '<~15%'. The main text (Section 5.4) gives '~3-4%' for SNR>=3 and '10-13%' for SNR=1. Consider making the abstract consistent with the more precise numbers in the text.

Circularity Check

0 steps flagged

No significant circularity: the k_noise calibration is an empirical fit to simulation data, not a self-definitional or self-citation-forced derivation

full rationale

The paper's derivation chain is self-contained and not circular. The Brunt method (Eqs. 5–11) is an independent, pre-existing method from Brunt et al. (2010a) that reconstructs 3D density dispersion from 2D column density power spectra. The paper's new contribution is the denoised estimate (Eq. 15), which truncates the Brunt sum at k_noise. The k_noise model (Eqs. 13–14) is an empirical fit calibrated on simulation data spanning three void sizes and five analysis regions. While the denoised estimate is then evaluated on the same simulation suite, this is standard empirical calibration and validation, not circular derivation: k_noise is defined as the intersection of the signal and noise spectra (a physical, measurable quantity), and the functional form (Eq. 13) is an empirical fit to that intersection — not defined in terms of the target output σ_ρ/ρ0. The noise-subtraction method (§5.1) is fully independent. The Brunt et al. (2010a) citation is external and independently verified. The only mild concern is that calibration and validation use the same simulation family, but this is a generalization risk (acknowledged in §5.3 and §6), not circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 3 axioms · 0 invented entities

The paper introduces no new physical entities or particles. The free parameters are all empirical fit constants for the k_noise model. The key axiom is the Brunt approximation (Eq. 9), which is a domain assumption from prior literature.

free parameters (4)
  • SNR_crit = (0.82 ± 0.03) * k_max^(2/3)
    Fitted to simulation data in Eq. (14) and Fig. 8. This is the characteristic SNR scale in the k_noise model.
  • Exponent -8/3 in Eq. (13) = -8/3
    Empirical fit exponent for the k_noise/SNR relation. Not derived from first principles.
  • Exponent 1/4 in Eq. (13) = 1/4
    Empirical fit exponent for the k_noise/SNR relation. Not derived from first principles.
  • Exponent 2/3 in Eq. (14) = 2/3
    Empirical fit exponent for the SNR_crit vs k_max relation. Not derived from first principles.
axioms (3)
  • domain assumption P3D(ρ,k) ≈ 2k * P2D(N,k) (Brunt approximation, Eq. 9)
    The core assumption of the Brunt method, invoked in §2.5, that the 2D column density spectrum can be extended to approximate the 3D density spectrum. This is not re-derived but taken from Brunt et al. (2010a).
  • domain assumption The signal power spectrum is a decreasing function of k
    Invoked in §5.2 to justify the existence of a signal-noise intersection point k_noise. Stated as typical of turbulent density structures.
  • ad hoc to paper k_noise is largely insensitive to Mach number, driving mode, and magnetic field
    Invoked in §5.3 and §6 to justify the universality of Eqs. (13)-(14). The authors note this may not hold at much higher Mach number but use it as a working assumption.

pith-pipeline@v1.1.0-glm · 19305 in / 2667 out tokens · 362294 ms · 2026-07-09T17:17:20.018566+00:00 · methodology

0 comments
read the original abstract

Turbulence plays a central role in shaping the structure and dynamics of the interstellar medium (ISM), governing the star formation rate (SFR) and the initial mass function (IMF). A key consequence of turbulence is the generation of density fluctuations, which regulate the amount of dense gas available for star formation. Accurate measurements of the three-dimensional (3D) turbulent density dispersion are therefore essential for understanding molecular-cloud structure and star formation. However, observations typically provide only two-dimensional (2D) column densities and are often affected by measurement/detector noise. The Brunt method estimates the 3D density dispersion from 2D column-density maps, but it does not account for finite signal-to-noise ratio (SNR). Here, we extend the method to recover the 3D turbulent density dispersion from noise-contaminated observations. Using numerical simulations spanning a range of density perturbation amplitudes and noise types, we identify a characteristic noise wavenumber, k_noise, corresponding to the intersection of the signal and noise spectra. Restricting the Brunt reconstruction to wavenumbers below k_noise yields a denoised density-dispersion estimate that closely reproduces the noise-free result. We provide a practical prescription to determine k_noise directly from the measurement SNR and image resolution. Alternatively, if the noise spectrum is known, it can be subtracted directly from the observed spectrum, eliminating the need to estimate k_noise. The proposed correction recovers the noise-free density dispersion with errors of <~5% for SNR>=3 and <~15% for SNR>=1, enabling substantially more reliable estimates of turbulent density fluctuations from noisy column-density data.

Figures

Figures reproduced from arXiv: 2607.07205 by Christoph Federrath, David C. Collins, Luz L. Jimenez Vela, Seth Davidovits.

Figure 1
Figure 1. Figure 1: Locations of the turbulence analysis regions (marked as white squares) in the shocktube simulation with 50 µm foam voids from Dhawalikar et al. (2022). Density slices at z = 0 (top panel) and y = 1.5 mm (bottom panel) are shown at t = 75 ns, i.e., just as the shock wave is ex￾iting the tube at y ∼ 1.8 mm, leaving behind a turbulent post-shock region. Five different turbulence analysis regions are selected … view at source ↗
Figure 2
Figure 2. Figure 2: Density projections along the y-axis of the center region with different levels of simulated white noise added – from left to right: SNR = 0.1, 1, 3, and 10. Noise dominates over the signal for SNR ≪ 1, while a nearly perfect signal is provided when SNR ≫ 1. Cases with SNR ∼ 1 are most interesting in that we can still reasonably recover the signal, but it may be strongly affected by noise, which we aim to … view at source ↗
Figure 3
Figure 3. Figure 3: Brunt density dispersion estimate σρ/ρ0,Brunt as a function of SNR. The data points show the median and 16th to 84th percentile range over the 15 samples (5 positions with 3 projections each). As SNR → ∞, the density dispersion reconstruction via the Brunt method works as expected, ap￾proaching the intrinsic σρ/ρ0 (shown as the dashed horizontal line with shaded area) to within ∼ 20% accuracy, while for SN… view at source ↗
Figure 6
Figure 6. Figure 6: shows knoise normalized to kmax as a function of SNR for the 12.5, 50, and 100 µm void size cases (the respective spectra for the 12.5 and 100 µm void size cases 10−1 100 101 102 SNR 10−1 100 knoise/kmax 12.5 µm 50 µm 100 µm [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Same as [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The SNRcrit in Eq. (13) as a function of kmax for each of the three void-size cases. The line is a fit following Eq. (14). allows for higher spatial resolution, and therefore in￾creased kmax, is also subject to increased noise on the smallest scales. At fixed SNR, this means that the noise spectra shift down when kmax increases, resulting in a change in the intersection wavenumber knoise. We quan￾tify this… view at source ↗
Figure 9
Figure 9. Figure 9: Same as [PITH_FULL_IMAGE:figures/full_fig_p009_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Same as [PITH_FULL_IMAGE:figures/full_fig_p010_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Same as [PITH_FULL_IMAGE:figures/full_fig_p010_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Same as [PITH_FULL_IMAGE:figures/full_fig_p011_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Top panel: similar to [PITH_FULL_IMAGE:figures/full_fig_p011_13.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

42 extracted references · 42 canonical work pages · 5 internal anchors

  1. [1]

    M., & Ryu , D

    Akahori , T., Gaensler , B. M., & Ryu , D. 2014, , 790, 123, 10.1088/0004-637X/790/2/123

    work page doi:10.1088/0004-637x/790/2/123 2014

  2. [2]

    Brunt , C. M. 2010, , 513, A67, 10.1051/0004-6361/200913506

    work page doi:10.1051/0004-6361/200913506 2010

  3. [3]

    M., & Federrath , C

    Brunt , C. M., & Federrath , C. 2014, , 442, 1451, 10.1093/mnras/stu888

    work page doi:10.1093/mnras/stu888 2014

  4. [4]

    2009, title Clustering of luminous red galaxies - III

    Brunt , C. M., Federrath , C., & Price , D. J. 2010 a , , 403, 1507, 10.1111/j.1365-2966.2009.16215.x

    work page doi:10.1111/j.1365-2966.2009.16215.x 2010

  5. [5]

    doi:10.1111/j.1745-3933.2010.00930.x , journal=

    ---. 2010 b , , 405, L56, 10.1111/j.1745-3933.2010.00858.x

    work page doi:10.1111/j.1745-3933.2010.00858.x 2010

  6. [6]

    2012, , 755, L19, 10.1088/2041-8205/755/1/L19

    Burkhart , B., & Lazarian , A. 2012, , 755, L19, 10.1088/2041-8205/755/1/L19

    work page doi:10.1088/2041-8205/755/1/l19 2012

  7. [7]

    2022, , 105, 065206, 10.1103/PhysRevE.105.065206

    Davidovits , S., Federrath , C., Teyssier , R., et al. 2022, , 105, 065206, 10.1103/PhysRevE.105.065206

    work page doi:10.1103/physreve.105.065206 2022

  8. [8]

    2022, , 514, 1782, 10.1093/mnras/stac1480

    Dhawalikar , S., Federrath , C., Davidovits , S., et al. 2022, , 514, 1782, 10.1093/mnras/stac1480

    work page doi:10.1093/mnras/stac1480 2022

  9. [9]

    2008, in Astronomical Society of the Pacific Conference Series, Vol

    Dubey , A., Fisher , R., Graziani , C., et al. 2008, in Astronomical Society of the Pacific Conference Series, Vol. 385, Numerical Modeling of Space Plasma Flows, ed. N. V. Pogorelov , E. Audit , & G. P. Zank , 145

    work page 2008

  10. [10]

    Elmegreen , B. G. 2009, in IAU Symposium, Vol. 254, IAU Symposium, ed. J. Andersen, J. Bland-Hawthorn, & B. Nordstr \"o m , 289, 10.1017/S1743921308027713

    work page internal anchor Pith review doi:10.1017/s1743921308027713 2009

  11. [11]

    Thermodynamics and dynamics of two-dimensional systems with dipole-like repulsive interactions

    Elmegreen , B. G., & Scalo , J. 2004, , 42, 211, 10.1146/annurev.astro.41.011802.094859

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1146/annurev.astro.41.011802.094859 2004

  12. [12]

    Federrath , C., & Klessen , R. S. 2012, , 761, 156, 10.1088/0004-637X/761/2/156

    work page doi:10.1088/0004-637x/761/2/156 2012

  13. [13]

    2013, , 763, 51, 10.1088/0004-637X/763/1/51

    ---. 2013, , 763, 51, 10.1088/0004-637X/763/1/51

    work page doi:10.1088/0004-637x/763/1/51 2013

  14. [14]

    S., Iapichino , L., & Beattie , J

    Federrath , C., Klessen , R. S., Iapichino , L., & Beattie , J. R. 2021, Nature Astronomy, 5, 365, 10.1038/s41550-020-01282-z

    work page doi:10.1038/s41550-020-01282-z 2021

  15. [15]

    S., & Schmidt , W

    Federrath , C., Klessen , R. S., & Schmidt , W. 2008, , 688, L79, 10.1086/595280

    work page doi:10.1086/595280 2008

  16. [16]

    2009, , 692, 364, 10.1088/0004-637X/692/1/364

    ---. 2009, , 692, 364, 10.1088/0004-637X/692/1/364

    work page doi:10.1088/0004-637x/692/1/364 2009

  17. [17]

    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, , 512, A81, 10.1051/0004-6361/200912437

    work page doi:10.1051/0004-6361/200912437 2010

  18. [18]

    M., Medling , A

    Federrath , C., Salim , D. M., Medling , A. M., et al. 2017, , 468, 3965, 10.1093/mnras/stx727

    work page doi:10.1093/mnras/stx727 2017

  19. [19]

    and Olson, K

    Fryxell , B., Olson , K., Ricker , P., et al. 2000, , 131, 273, 10.1086/317361

    work page doi:10.1086/317361 2000

  20. [20]

    P., & Klessen , R

    Girichidis , P., Konstandin , L., Whitworth , A. P., & Klessen , R. S. 2014, , 781, 91, 10.1088/0004-637X/781/2/91

    work page doi:10.1088/0004-637x/781/2/91 2014

  21. [21]

    E., Lee , M

    Hamilton , C. E., Lee , M. N., & Parra-Vasquez , A. N. G. 2016, Fusion Science and Technology, 70, 226, 10.13182/FST15-227

    work page doi:10.13182/fst15-227 2016

  22. [22]

    Hew , J. K. J., & Federrath , C. 2023, , 520, 6268, 10.1093/mnras/stad545

    work page doi:10.1093/mnras/stad545 2023

  23. [23]

    2014, Science, 344, 183, 10.1126/science.1248724

    Kainulainen , J., Federrath , C., & Henning , T. 2014, Science, 344, 183, 10.1126/science.1248724

    work page doi:10.1126/science.1248724 2014

  24. [24]

    2005, , 630, L45, 10.1086/491600

    Kim , J., & Ryu , D. 2005, , 630, L45, 10.1086/491600

    work page doi:10.1086/491600 2005

  25. [25]

    S., et al

    Kitsionas , S., Federrath , C., Klessen , R. S., et al. 2009, , 508, 541, 10.1051/0004-6361/200811170

    work page doi:10.1051/0004-6361/200811170 2009

  26. [26]

    S., & Schmidt , W

    Konstandin , L., Federrath , C., Klessen , R. S., & Schmidt , W. 2012, Journal of Fluid Mechanics, 692, 183, 10.1017/jfm.2011.503

    work page internal anchor Pith review doi:10.1017/jfm.2011.503 2012

  27. [27]

    2007, , 658, 423, 10.1086/511515

    Kowal , G., Lazarian , A., & Beresnyak , A. 2007, , 658, 423, 10.1086/511515

    work page doi:10.1086/511515 2007

  28. [28]

    G., Norman , M

    Kritsuk , A. G., Norman , M. L., & Wagner , R. 2011 a , , 727, L20, 10.1088/2041-8205/727/1/L20

    work page doi:10.1088/2041-8205/727/1/l20 2011

  29. [29]

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

    Kritsuk , A. G., Nordlund , A ., Collins , D., et al. 2011 b , , 737, 13, 10.1088/0004-637X/737/1/13

    work page doi:10.1088/0004-637x/737/1/13 2011

  30. [30]

    Larson , R. B. 1981, , 194, 809, 10.1093/mnras/194.4.809

    work page doi:10.1093/mnras/194.4.809 1981

  31. [31]

    Mac Low , M.-M., & Klessen , R. S. 2004, Reviews of Modern Physics, 76, 125, 10.1103/RevModPhys.76.125

    work page doi:10.1103/revmodphys.76.125 2004

  32. [32]

    S., Federrath , C., & Seta , A

    Mathew , S. S., Federrath , C., & Seta , A. 2023, , 518, 5190, 10.1093/mnras/stac3415

    work page doi:10.1093/mnras/stac3415 2023

  33. [33]

    Malicious User Experience Design Research for Cybersecurity

    McKee , C. F., & Ostriker , E. C. 2007, , 45, 565, 10.1146/annurev.astro.45.051806.110602

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1146/annurev.astro.45.051806.110602 2007

  34. [34]

    R., Raman , K

    Nagel , S. R., Raman , K. S., Huntington , C. M., et al. 2017, Physics of Plasmas, 24, 072704, 10.1063/1.4985312

    work page doi:10.1063/1.4985312 2017

  35. [35]

    Padoan , P., Nordlund , A., & Jones , B. J. T. 1997, , 288, 145, 10.1093/mnras/288.1.145

    work page doi:10.1093/mnras/288.1.145 1997

  36. [36]

    1998, , 58, 4501, 10.1103/PhysRevE.58.4501

    Passot , T., & V \'a zquez-Semadeni , E. 1998, , 58, 4501, 10.1103/PhysRevE.58.4501

    work page doi:10.1103/physreve.58.4501 1998

  37. [37]

    MNRAS , author =

    Price , D. J., & Federrath , C. 2010, , 406, 1659, 10.1111/j.1365-2966.2010.16810.x

    work page doi:10.1111/j.1365-2966.2010.16810.x 2010

  38. [38]

    J., Federrath , C., & Brunt , C

    Price , D. J., Federrath , C., & Brunt , C. M. 2011, , 727, L21, 10.1088/2041-8205/727/1/L21

    work page doi:10.1088/2041-8205/727/1/l21 2011

  39. [39]

    J., & P \'e rez , E

    S \'a nchez , N., Alfaro , E. J., & P \'e rez , E. 2005, , 625, 849, 10.1086/429553

    work page doi:10.1086/429553 2005

  40. [40]

    1998, , 336, 697

    Stutzki , J., Bensch , F., Heithausen , A., Ossenkopf , V., & Zielinsky , M. 1998, , 336, 697

    work page 1998

  41. [41]

    2011, Journal of Computational Physics, 230, 3331, 10.1016/j.jcp.2011.01.026

    Waagan , K., Federrath , C., & Klingenberg , C. 2011, Journal of Computational Physics, 230, 3331, 10.1016/j.jcp.2011.01.026

    work page internal anchor Pith review doi:10.1016/j.jcp.2011.01.026 2011

  42. [42]

    2024, , 971, 48, 10.3847/1538-4357/ad5a84

    Yoon , H., & Cho , J. 2024, , 971, 48, 10.3847/1538-4357/ad5a84

    work page doi:10.3847/1538-4357/ad5a84 2024