Pith. sign in

REVIEW 2 major objections 5 minor 106 references

The Density Distribution of Compressively-Forced Supersonic Turbulence Depends on the Driving Correlation Time

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read This paper claims that the correlation time of the driving accelerations is a third control parameter of the density distribution in compressively-driven supersonic turbulence, at fixed Mach number and driving mix.

desk verdict A genuinely new control parameter for the density PDF in compressive supersonic turbulence, with a large effect that is well resolved; the missing error bars and seed policy are the main soft spots, but the paper deserves peer review. read the letter →

arxiv 2505.23898 v1 pith:S7S4777Q submitted 2025-05-29 astro-ph.GA

classification astro-ph.GA
keywords supersonicturbulencedensityPDFcompressiveforcingdrivingcorrelationtimeOrnstein-UhlenbeckprocessinterstellarmediumLagrangiantracerparticlesmolecularclouds
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

The paper sets out to show that the density distribution of compressively-driven supersonic turbulence is not set by the Mach number and the mix of compressive versus solenoidal driving alone: the correlation time of the driving accelerations, $\tau_a$, is a third control parameter. In simulations with $1024^3$ cells and $512^3$ tracer particles, the authors find that when $\tau_a$ is comparable to the eddy turnover time $\tau_e$, compressive forcing sustains accelerated expansions that carve out large low-density voids ringed by shocks, producing a broad, negatively skewed density PDF. When $\tau_a$ is shortened to about $0.1\tau_e$, the voids nearly disappear and the PDF narrows: in the purely compressive runs at comparable Mach number, the volume-weighted density variance $\sigma^2_{s,V}$ drops from $9.07$ to $5.37$. The same change in $\tau_a$ leaves solenoidally-driven turbulence essentially untouched. A sympathetic reader would care because density statistics feed directly into predictions of star formation rates and observed column-density structure in the interstellar medium.

What carries the argument

The load-bearing object is the Ornstein-Uhlenbeck (OU) driving process of eq. (4), a Fourier-space random acceleration with a single correlation time $\tau_a$ and a fixed spectral shape, combined with the projection tensor of eq. (5) that splits the forcing into solenoidal and compressive parts. The paper's diagnostics are Lagrangian tracer particles ($512^3$ per run, cloud-in-cell interpolated) that measure $s$ and its time derivatives along fluid paths, plus the identity $ds/dt = -\nabla\cdot v$ from the continuity equation. The argument runs through eq. (10), which expresses the stationary mass-weighted PDF of $s$ entirely in terms of the conditional averages $\langle(ds/dt)^2|s\rangle$ and $\langle d^2s/dt^2|s\rangle$, and eq. (11), the decomposition of $d^2s/dt^2$ into nonlinear, pressure, viscous, and driving terms. The driving term $\nabla\cdot a$ is small in magnitude but acts as a coherent seed whose nonlinear amplification is controlled by $\tau_a$; that amplification mechanism, operating in slowly-evolving low-density regions, is what carries the dependence of the density PDF on the correlation time.

What would settle it

Run a purely compressive turbulence simulation at fixed Mach number and effective viscosity with $\tau_a/\tau_e$ varied in fine steps from roughly $0.01$ to $3$: if the volume-weighted density variance and skewness do not fall monotonically toward the short-$\tau_a$ values measured here ($\sigma^2_{s,V} = 5.37$, $\mu_{s,V} \approx -0.17$), the central claim fails. A complementary test in a supernova-driven ISM simulation is to measure the correlation coefficient between $s$ and $\nabla\cdot a$: the mechanism predicts a substantially negative coefficient wherever sustained voids form, and a value near zero if the feedback acts as short-correlation driving.

Watch

Extended reading notes

Core claim

The central claim is that the logarithmic density PDF in compressively-forced supersonic turbulence depends on how long the forcing stays coherent, not only on its strength and geometry. Concretely: with purely compressive driving at Mach number around 7, reducing $\tau_a/\tau_e$ from $0.51$ to $0.06$ reduces the volume-weighted variance $\sigma^2_{s,V}$ from $9.07$ to $5.37$ and roughly halves the negative skewness of $P_M(s)$ (Table 2). The paper traces this to the divergence of the driving acceleration field, $\nabla\cdot a$. Because the forcing is applied on large scales, $\nabla\cdot a$ directly seeds $ds/dt = -\nabla\cdot v$; when $\tau_a$ is long, the seed stays coherent long enough that the nonlinear term in the $d^2s/dt^2$ equation amplifies it preferentially in low-density regions, where $ds/dt$ evolves slowly. The result is sustained accelerated expansion: large voids that are slowly expanding in their interiors and bounded by shocks, a large conditional variance $\langle(ds/dt)^2|s\rangle$ at low $s$, and, through the identity $P_M(s) \propto \langle(ds/dt)^2|s\rangle^{-1}\exp\left(\int ds'\, \langle d^2s/dt^2|s'\rangle\big/\langle(ds/dt)^2|s'\rangle\right)$, a broader and more negatively skewed mass-weighted PDF. For solenoidal driving $\nabla\cdot a = 0$, so no such amplification path exists and $\tau_a$ has negligible effect.

Load-bearing premise

The results assume that real compressive astrophysical driving (for example, supernova feedback) can be represented by a single-scale random forcing with one correlation time and a fixed spectral shape; if real driving injects momentum in localized bursts spread over a range of timescales, the quantitative dependence of the density PDF on a single $\tau_a$ may not transfer.

Editorial extensions

If this is right

  • Simulations of compressively-driven turbulence that omit $\tau_a/\tau_e$ as a parameter will over- or under-predict the low-density tail of the density distribution at fixed Mach number.
  • Volume-weighted statistics are far more sensitive to $\tau_a$ than mass-weighted ones: $\sigma^2_{s,V}$ changes by roughly a factor of 1.7 between the long and short runs while $\sigma^2_{s,M}$ changes by only a few percent, so comparisons with column-density observations should report the driving correlation time.
  • In supernova-driven interstellar turbulence, simple momentum-conservation estimates give $\tau_a/\tau_e \approx 0.25$ or smaller, placing real galactic driving between the two regimes studied here and implying that the density structure of the ISM depends on the timing of feedback, not just its total energy.
  • The conditional-variance mechanism identified here, with $\langle(ds/dt)^2|s\rangle$ rising at low $s$ because of seeded, amplified expansions, provides a direct test of whether any given simulation's driving prescription produces physically faithful density PDFs.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension the paper does not run: vary $\tau_a/\tau_e$ continuously from roughly $0.01$ to several at fixed Mach number and viscosity in purely compressive runs; if the variance and skewness saturate at both ends rather than changing monotonically, the two end states measured here bracket a smooth transition that a one-parameter formula could capture.
  • Real supernova feedback is bursty, localized, and multi-scale; if it has a spectrum of correlation times rather than a single $\tau_a$, its effective coherence is likely shorter than the eddy time on all but the largest scales, which would push real ISM density PDFs toward the narrow end of the range measured here.
  • Because the mechanism relies on $\nabla\cdot a$ staying coherent for roughly an eddy time while correlated with low-density regions, adding a long-range compressive component such as self-gravity to an otherwise solenoidal simulation should widen the low-density tail in the same way, a prediction that gravoturbulent runs could test.
  • Observational consequence: at comparable Mach number, regions driven by clustered, quasi-steady feedback should show broader column-density PDFs than regions driven by short, stochastic bursts, so comparing N-PDFs across star-forming regions with matched Mach numbers could indirectly constrain the effective driving correlation time.
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

2 major / 5 minor

Summary. This paper investigates whether the autocorrelation time of the driving acceleration, τ_a, affects the density PDF of isothermal supersonic turbulence. The authors run 1024^3 simulations with explicit viscosity and ILES variants, plus 512^3 ILES checks, with solenoidal, mixed, and compressive forcing, comparing a long correlation time (τ_a ≈ 0.6τ_e) with a short one (τ_a ≈ 0.06τ_e). They find that for solenoidal forcing the density statistics are nearly unchanged, while for compressive forcing the long-τ_a runs develop large low-density voids bounded by shocks, producing a broader and more negatively skewed volume-weighted density PDF (σ_s,V^2 = 9.07 versus 5.37 in the fiducial viscous runs). Using Lagrangian tracers, they measure conditional averages of ds/dt and d^2s/dt^2 and show that the enhancement of the conditional variance ⟨(ds/dt)^2|s⟩ at low s, driven by a coherent ∇·a seed amplified by the nonlinear dynamics, explains the broader PDF. They close with an analytic estimate suggesting τ_a/τ_e ≈ 0.25 in a supernova-driven ISM, between the two simulated limits.

Significance. If the central claim holds, this work establishes that the density PDF of compressively forced supersonic turbulence is not determined by Mach number and driving mix alone; the driving correlation time is a third parameter with potentially large effects. That matters for modeling the density structure of molecular clouds and the ISM and for interpreting simulations that use different driving decorrelation times. The paper's strengths are the direct comparison of simulations without fitted parameters, the consistency checks across explicit-viscosity DNS, ILES, and lower-resolution runs, the use of tracer particles to build a mechanistic explanation, the reconstruction of PM(s) from measured conditional averages via Eq. (10), and the unusually transparent numerical setup (fixed code commit, tabulated run parameters).

major comments (2)
  1. [§3.3, Table 2] The headline quantitative result—σ_s,V^2 dropping from 9.07 to 5.37 when τ_a/τ_e changes from 0.51 to 0.06—is reported without any uncertainty estimate. The values are time averages over the 3–7 τ_e stationary window; for the L runs τ_a ≈ 0.5τ_e, so that window contains only about eight independent forcing decorrelation times, and σ_s,V^2 is dominated by a small number of large voids. The agreement among the viscous, ILES, and 512^3 runs is supportive, but if those runs share forcing seeds, they are not independent realizations. Please add per-statistic uncertainties (for example block or bootstrap estimates over independent subintervals), a stationarity test for the averaging window, and a statement of the seed policy. If the run-to-run scatter is comparable to the Δσ^2 ≈ 3.7 gap, the quantitative magnitude of the effect is not established, even though the qualitative direction appears robust.
  2. [§3.3, Table 2] The claim that the two compressive runs being compared have 'comparable Mach number' should be made precise. In the fiducial pair carrying the main comparison, the volume-weighted Mach numbers are indeed close (7.3 versus 7.4), but the mass-weighted Mach numbers differ substantially (5.6 versus 7.1). Because σ_s^2 depends on Mach number, this mismatch could in principle contaminate the comparison. The direction of the observed effect is opposite to what a higher mass-weighted Mach would naively produce, so the concern is likely not fatal, but the paper should state this explicitly and, if possible, demonstrate the τ_a dependence at more closely matched mass-weighted Mach numbers or otherwise control for the Mach dependence.
minor comments (5)
  1. [Abstract and §4] The abstract contains the grammatical error 'τa is may be significantly less'; this should be corrected to 'τa may be significantly less.'
  2. [§2.1, Eq. (4)] The definition of the forcing spectrum Pa(k) = k^2(2−k^2)Θ(k^2−2) appears to have the Heaviside argument reversed; to make the profile positive and peaked at k = kp, the argument should be Θ(2−k^2). Please check the intended expression.
  3. [Fig. 3 caption] The caption says the black lines are '∝ (−ds/dt)^−6 on the left and ∝ (−ds/dt)^−3'; the second power law should be for positive ds/dt, i.e., ∝ (ds/dt)^−3 on the right, consistent with the text in §3.3.
  4. [Table 3] The caption states that Table 3 lists the mean and variance of 'PV(ds/dt)' computed from the particles, but the column labels and the discussion indicate these are mass-weighted statistics; the PV label should be PM.
  5. [§4] The symbol τcorr appears without definition in the sentence 'if Rmom ≈ L, then τcorr is shorter'; this should be τ_a to match the rest of the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claim is a direct comparison of simulation outputs at different input forcing correlation times, with no fitted parameter or self-citation chain forcing the result.

full rationale

The load-bearing claim is that for purely compressive driving the volume-weighted density variance drops from sigma^2_s,V = 9.07 to 5.37 when tau_a/tau_e is changed from 0.51 to 0.06 (Table 2). This is a direct comparison of simulation outputs in which tau_a is an input parameter of the Ornstein-Uhlenbeck driving scheme (eq. 4, via c_drift = e^{-Delta t/tau_a}), and no parameter is fitted to the density PDF to produce the difference. The projection tensor (eq. 5) and forcing spectrum are also prescribed inputs, independent of the measured density statistics. The conditional-average framework in eqs. (8)-(10) is used as a steady-state consistency check: eq. (10) is an exact identity from the Fokker-Planck / Pope-Ching formalism, and the paper compares the PM(s) reconstructed from measured conditional averages with the directly measured PM(s) (Fig. 9, right panel). That is confirmation of stationarity and tracer accuracy, not a prediction derived from the effect being claimed. Citations to Scannapieco et al. (2024) supply the tracer methodology and earlier solenoidal-driving results, but the compressive tau_a dependence is not imported from those papers; the cited equation from Pan et al. (2019) is an exact kinematic identity that does not contain the claimed tau_a dependence. The supernova correlation-time estimate in Sec. 4 is a separate analytic argument from energy and momentum balance. Concerns about run-to-run scatter, seed independence, or the single-realization time average are statistical robustness issues, not circularity.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

No free parameters are fitted to produce the central result; the listed parameters are numerical settings (viscosity) and a constant in the rough astrophysical extrapolation. No new physical entities are introduced.

free parameters (2)
  • explicit viscosity nu = 5.5e-4 (box units)
    Chosen to match the effective viscosity of the 512^3 ILES runs (Sec. 2.2); the central tau_a comparison is robust to this choice because the 1024^3 ILES runs show the same trend (Table 3 and Appendix Fig. 8).
  • fSN = order unity
    A multiplicative constant in the analytic supernova-driving estimate (eq. 12, Sec. 4); it affects the derived tau_a/tau_e ~ 0.25 but is not used in the simulation-based central claim.
assumptions (3)
  • domain assumption The isothermal Euler equations with stochastic forcing capture the essential physics of supersonic turbulence relevant to the density PDF.
    Invoked throughout Sec. 2.1; the paper does not include magnetic fields, gravity, or cooling, so the quantitative results may not carry over.
  • domain assumption The Ornstein-Uhlenbeck process with a single correlation time tau_a and fixed spectral shape Pa(k) is an adequate model of time-dependent compressive driving.
    Introduced in Sec. 2.1, eq. (4); the central claim is framed in terms of tau_a, so if real driving has a spectrum of correlation times, the 1D tau_a parameter may not capture it.
  • standard math Equation (10) (from Pope & Ching 1993; Pan et al. 2019) relates the steady-state PM(s) to conditional averages of ds/dt and d^2s/dt^2.
    Used in Sec. 3.4 to interpret the conditional statistics; the paper confirms it reproduces the measured PM(s) in Appendix Fig. 9.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The Density Distribution of Compressively-Forced Supersonic Turbulence Depends on the Driving Correlation Time." pith.science (2026). https://pith.science/paper/S7S4777Q

@misc{pith2026250523898,
  author       = {Pith},
  title        = {Pith review of: The Density Distribution of Compressively-Forced Supersonic Turbulence Depends on the Driving Correlation Time},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S7S4777Q}},
  note         = {Machine review of arXiv:2505.23898}
}
abstract

Supersonic turbulence plays a critical role in shaping astrophysical systems, from molecular clouds to the circumgalactic medium. Key properties of this turbulence include the Mach number, driving scale, and nature of the driving mechanism, which can be solenoidal (divergence-free), compressive (curl-free), or a mix of the two. A less studied property is the correlation time of the driving accelerations, $\tau_{\rm a}.$ While this timescale has a minimal impact on solenoidally-driven turbulence, we show that it has a strong impact on compressively-driven turbulence. Using high-resolution simulations with tracer particles, we analyze the evolution of density fluctuations, focusing on the PDF of the logarithmic density, $s$, and its rate of change, $\frac{ds}{dt},$ and the conditional statistics of $\frac{ds}{dt}$ and $\frac{d^2s}{dt^2}$. When the driving correlation time is comparable to the eddy turnover time, $\tau_{\rm a} \approx \tau_{\rm e},$ compressive driving leads to the formation of large, low-density voids in which the variance of $\frac{ds}{dt}$ is large. These are directly linked to sustained accelerated expansions, which results in a strong correlation between density and the divergence of the driving acceleration field. In contrast, when $\tau_{\rm a} \approx 0.1 \, \tau_{\rm e}$, compressive driving does not produce such voids, resulting in a narrower, less skewed distribution. We show using analytical estimates that $\tau_{\rm a}$ is may be significantly less than $\tau_{\rm e}$ in supernova-driven turbulence, highlighting the need to better understand the role of the driving correlation time in shaping the density structure of turbulent astrophysical systems.

Figures

Figures reproduced from arXiv: 2505.23898 by the authors.

Figure 1
Figure 1. Compensated specific kinetic energy spectra from simulations with explicit viscosity, showing total (left), solenoidal (center), and compressive components (right) versus normalized wavenumber. Each panel displays results from runs 0.0-Lν (orange solid), 0.0-sν (red dashed), 0.3-Lν (magenta solid), 0.3-sν (purple dashed), 1.0-Lν (green solid), and 1.0-sν (blue dashed). Lines represent mean spectra over the stationar… view at source ↗
Figure 2
Figure 2. Representative results from our turbulence simulations. From left to right, columns show results from runs with purely solenoidal, mixed driving, and purely compressive driving. From top to bottom, rows show the logarithm of the projected density through the box, and slices of s, ∇·u, and ∇·a. For consistency, each column shows results from the snapshot with the highest Mach number from the corresponding simulations… view at source ↗
Figure 4
Figure 4. Ratio of the average value of ds dt as a function of s nor￾malized by σds/dt,M. The lines are as above. This quantity should be zero in a steady state, and it thus serves as a test of the accuracy of our measurements (Pan et al. 2018). out of each s bin to be zero (Pan et al. 2018). Intuitively, the probability flux is given by PM(s) [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figures from the paper (6 more)
Figure 3
Figure 3. Figure 3: Top: Mass-weighted PDF of s. As in [PITH_FULL_IMAGE:figures/full_fig_p008_3.png]
Figure 5
Figure 5. Figure 5: Left: The average value of [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Divergence of the acceleration field, normalized to be directly comparable to the right panel of [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Two-dimensional volume-weighted PDF of normalized log density s−⟨s⟩ v and normalized divergence of the acceleration field ∆x∇· a/arms. Each panel provides the correlation coefficient between the two quantities. Only for large τa values in combination with compressive d…
Figure 8
Figure 8. Figure 8: Left: Specific kinetic energy spectra comparing different numerics (10243 DNS, 10243 ILES, and 5123 ILES) at the extreme points in our parameter space, i.e., for ζ = 0.0 and long τ a, and ζ = 1.0 and short τ a. The ζ = 0.0 spectra are vertically offset for better visib…
Figure 9
Figure 9. Figure 9: Left: The average value of [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

106 extracted references · 71 canonical work pages

  1. [1]

    602C `\.=

    thebibliography [1] 20pt to REFERENCES 6pt =0pt -12pt 10pt plus 3pt =0pt =0pt =1pt plus 1pt =0pt =0pt -12pt =13pt plus 1pt =20pt =13pt plus 1pt \@M =10000 =-1.0em =0pt =0pt 0pt =0pt =1.0em @enumiv\@empty 10000 10000 `\.\@m \@noitemerr \@latex@warning Empty `thebibliography' environment \@ifnextchar \@reference \@latexerr Missing key on reference command E...

  2. [2]

    1999, Physics of Fluids, 11, 1880

    Alvelius, K. 1999, Physics of Fluids, 11, 1880

  3. [3]

    2020, , 641, A70

    Bacchini , C., Fraternali , F., Iorio , G., et al. 2020, , 641, A70

  4. [4]

    S., Mac Low , M

    Ballesteros-Paredes , J., Klessen , R. S., Mac Low , M. M., & Vazquez-Semadeni , E. 2007, in Protostars and Planets V, ed. B. Reipurth , D. Jewitt , & K. Keil , 63

  5. [5]

    1996, , 34, 155

    Beck , R., Brandenburg , A., Moss , D., Shukurov , A., & Sokoloff , D. 1996, , 34, 155

  6. [6]

    M., Wright, E

    Blondin, J. M., Wright, E. B., Borkowski, K. J., & Reynolds, S. P. 1998, The Astrophysical Journal, 500, 342

  7. [7]

    F., & Torkelsson , U

    Brandenburg , A., Nordlund , A., Stein , R. F., & Torkelsson , U. 1995, , 446, 741

  8. [8]

    2025, The Astrophysical Journal, 983, 105

    Brandenburg, A., & Scannapieco, E. 2025, The Astrophysical Journal, 983, 105

Show all 106 references
  1. [9]

    M., & Mac Low , M.-M

    Brunt , C. M., & Mac Low , M.-M. 2004, , 604, 196

  2. [10]

    J., & Scannapieco , E

    Buie II , E., Gray , W. J., & Scannapieco , E. 2018, , 864, 114

  3. [11]

    J., Scannapieco , E., & Safarzadeh , M

    Buie II , E., Gray , W. J., Scannapieco , E., & Safarzadeh , M. 2020, , 896, 136

  4. [12]

    2009, , 693, 250

    Burkhart , B., Falceta-Gon c alves , D., Kowal , G., & Lazarian , A. 2009, , 693, 250

  5. [13]

    2012, , 755, L19

    Burkhart , B., & Lazarian , A. 2012, , 755, L19

  6. [14]

    2020, Galaxies, 8, 56

    Chamandy , L., & Shukurov , A. 2020, Galaxies, 8, 56

  7. [15]

    2007, , 661, 262

    Dib , S., Kim , J., V \'a zquez-Semadeni , E., Burkert , A., & Shadmehri , M. 2007, , 661, 262

  8. [16]

    L., & Bonnell , I

    Dobbs , C. L., & Bonnell , I. A. 2008, , 385, 1893

  9. [17]

    C., Trott, C

    Edwards, H. C., Trott, C. R., & Sunderland, D. 2014, Journal of Parallel and Distributed Computing, 74, 3202

  10. [18]

    G., & Burkert , A

    Elmegreen , B. G., & Burkert , A. 2010, , 712, 294

  11. [19]

    G., & Scalo , J

    Elmegreen , B. G., & Scalo , J. 2004, , 42, 211

  12. [20]

    Eswaran, V., & Pope, S. B. 1988, Computers and Fluids, 16, 257

  13. [21]

    2015, , 448, 3297

    Federrath , C., & Banerjee , S. 2015, , 448, 3297

  14. [22]

    Federrath , C., & Klessen , R. S. 2013, , 763, 51

  15. [23]

    S., & Schmidt , W

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

  16. [24]

    S., Schmidt, W., & Mac Low, M

    Federrath, C., Roman-Duval, J., Klessen, R. S., Schmidt, W., & Mac Low, M. M. 2010, Astronomy & Astrophysics, 512, A81

  17. [25]

    M., Longmore , S

    Federrath , C., Rathborne , J. M., Longmore , S. N., et al. 2016, , 832, 143

  18. [26]

    M., Genzel , R., Bouch \'e , N., et al

    F \"o rster Schreiber , N. M., Genzel , R., Bouch \'e , N., et al. 2009, , 706, 1364

  19. [27]

    2000, , 131, 273

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

  20. [28]

    2016, , 456, 3432

    Girichidis , P., Walch , S., Naab , T., et al. 2016, , 456, 3432

  21. [29]

    Glover , S. C. O., & Clark , P. C. 2012, , 421, 116

  22. [30]

    Glover , S. C. O., & Mac Low , M.-M. 2007, , 659, 1317

  23. [31]

    J., Krumholz , M

    Goldbaum , N. J., Krumholz , M. R., Matzner , C. D., & McKee , C. F. 2011, , 738, 101

  24. [32]

    J., Scannapieco, E., & Kasen, D

    Gray, W. J., Scannapieco, E., & Kasen, D. 2015, , 801, 107

  25. [33]

    W., & O'Shea , B

    Grete , P., Glines , F. W., & O'Shea , B. W. 2021, IEEE Transactions on Parallel and Distributed Systems, 32, 85

  26. [34]

    W., & Beckwith , K

    Grete , P., O'Shea , B. W., & Beckwith , K. 2018, , 858, L19

  27. [35]

    C., Miller, J

    Grete, P., Dolence, J. C., Miller, J. M., et al. 2023, The International Journal of High Performance Computing Applications, 0, 10943420221143775

  28. [36]

    Y., Guszejnov , D., Hopkins , P

    Grudi \'c , M. Y., Guszejnov , D., Hopkins , P. F., Offner , S. S. R., & Faucher-Gigu \`e re , C.-A. 2021, , 506, 2199

  29. [37]

    2009, , 702, 1428

    Hennebelle , P., & Chabrier , G. 2009, , 702, 1428

  30. [38]

    2011, , 743, L29

    ---. 2011, , 743, L29

  31. [39]

    2014, , 570, A81

    Hennebelle , P., & Iffrig , O. 2014, , 570, A81

  32. [40]

    Hopkins , P. F. 2013, , 430, 1653

  33. [41]

    F., Narayanan , D., & Murray , N

    Hopkins , P. F., Narayanan , D., & Murray , N. 2013, , 432, 2647

  34. [42]

    K., Klessen , R

    Jappsen , A. K., Klessen , R. S., Larson , R. B., Li , Y., & Mac Low , M. M. 2005, , 435, 611

  35. [43]

    Kahniashvili , T., Brandenburg , A., Campanelli , L., Ratra , B., & Tevzadze , A. G. 2012, , 86, 103005

  36. [44]

    2011, , 530, A64

    Kainulainen , J., Beuther , H., Banerjee , R., Federrath , C., & Henning , T. 2011, , 530, A64

  37. [45]

    2013, , 553, L8

    Kainulainen , J., Federrath , C., & Henning , T. 2013, , 553, L8

  38. [46]

    2006, Journal of Turbulence, 7, 20

    Kaneda , Y., & Ishihara , T. 2006, Journal of Turbulence, 7, 20

  39. [47]

    Kim , C.-G., & Ostriker , E. C. 2017, , 846, 133

  40. [48]

    Kim , J.-G., Kim , W.-T., & Ostriker , E. C. 2018, , 859, 68

  41. [49]

    C., & Filippova , N

    Kim , J.-G., Ostriker , E. C., & Filippova , N. 2021, , 911, 128

  42. [50]

    Kim , W.-T., & Ostriker , E. C. 2007, , 660, 1232

  43. [51]

    C., & Stone , J

    Kim , W.-T., Ostriker , E. C., & Stone , J. M. 2003, , 599, 1157

  44. [52]

    S., & Hennebelle , P

    Klessen , R. S., & Hennebelle , P. 2010, , 520, A17

  45. [53]

    Konstandin , L., Girichidis , P., Federrath , C., & Klessen , R. S. 2012, , 761, 149

  46. [54]

    2023, , 956, 54

    Koplitz , B., Buie II , E., & Scannapieco , E. 2023, , 956, 54

  47. [55]

    2007, , 658, 423

    Kowal , G., Lazarian , A., & Beresnyak , A. 2007, , 658, 423

  48. [56]

    G., Padoan , P., Wagner , R., & Norman , M

    Kritsuk , A. G., Padoan , P., Wagner , R., & Norman , M. L. 2007, in American Institute of Physics Conference Series, Vol. 932, Turbulence and Nonlinear Processes in Astrophysical Plasmas, ed. D. Shaikh & G. P. Zank , 393--399

  49. [57]

    G., Ustyugov , S

    Kritsuk , A. G., Ustyugov , S. D., Norman , M. L., & Padoan , P. 2010, in Astronomical Society of the Pacific Conference Series, Vol. 429, Numerical Modeling of Space Plasma Flows, Astronum-2009, ed. N. V. Pogorelov , E. Audit , & G. P. Zank , 15

  50. [58]

    R., Matzner , C

    Krumholz , M. R., Matzner , C. D., & McKee , C. F. 2006, , 653, 361

  51. [59]

    R., & McKee, C

    Krumholz, M. R., & McKee, C. F. 2005, The Astrophysical Journal, 630, 250

  52. [60]

    Larson , R. B. 1981, , 194, 809

  53. [61]

    N., & Stone , J

    Lemaster , M. N., & Stone , J. M. 2008, , 682, L97

  54. [62]

    P., Cen , R., Bryan , G

    Li , M., Ostriker , J. P., Cen , R., Bryan , G. L., & Naab , T. 2015 a , , 814, 4

  55. [63]

    S., McKee , C

    Li , P. S., McKee , C. F., & Klein , R. I. 2015 b , , 452, 2500

  56. [64]

    2004, , 609, L83

    Li , Z.-Y., & Nakamura , F. 2004, , 609, L83

  57. [65]

    S., Kim , J., & de Avillez , M

    Mac Low , M.-M., Balsara , D. S., Kim , J., & de Avillez , M. A. 2005, , 626, 864

  58. [66]

    S., Burkert , A., & Smith , M

    Mac Low , M.-M., Klessen , R. S., Burkert , A., & Smith , M. D. 1998, , 80, 2754

  59. [67]

    2015, , 450, 504

    Martizzi , D., Faucher-Gigu \`e re , C.-A., & Quataert , E. 2015, , 450, 504

  60. [68]

    McKee , C. F. 1989, , 345, 782

  61. [69]

    D., Radice , D., et al

    M \"o sta , P., Ott , C. D., Radice , D., et al. 2015, Nature, 528, 376

  62. [70]

    K., & Padoan , P

    Nordlund , A . K., & Padoan , P. 1999, in Interstellar Turbulence, ed. J. Franco & A. Carraminana , 218

  63. [71]

    C., & Kim , C.-G

    Ostriker , E. C., & Kim , C.-G. 2022, , 936, 137

  64. [72]

    C., Stone , J

    Ostriker , E. C., Stone , J. M., & Gammie , C. F. 2001, , 546, 980

  65. [73]

    2011, , 730, 40

    Padoan , P., & Nordlund , . 2011, , 730, 40

  66. [74]

    Padoan , P., Nordlund , A., & Jones , B. J. T. 1997, , 288, 145

  67. [75]

    2016, The Astrophysical Journal, 822, 11

    Padoan , P., Pan , L., Haugb lle , T., & Nordlund , A . 2016, The Astrophysical Journal, 822, 11

  68. [76]

    2018, , 866, L17

    Pan , L., Padoan , P., & Nordlund , A . 2018, , 866, L17

  69. [77]

    2019, , 881, 155

    ---. 2019, , 881, 155

  70. [78]

    2010, , 721, 1765

    Pan , L., & Scannapieco , E. 2010, , 721, 1765

  71. [79]

    2022, arXiv e-prints, arXiv:2203.11179

    Pattle , K., Fissel , L., Tahani , M., Liu , T., & Ntormousi , E. 2022, arXiv e-prints, arXiv:2203.11179

  72. [80]

    S., & Mac Low , M.-M

    Peters , T., Banerjee , R., Klessen , R. S., & Mac Low , M.-M. 2011, , 729, 72

  73. [81]

    A., & Ostriker , E

    Piontek , R. A., & Ostriker , E. C. 2007, , 663, 183

  74. [82]

    B., & Ching , E

    Pope , S. B., & Ching , E. S. C. 1993, Physics of Fluids A, 5, 1529

  75. [83]

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

    Price , D. J., Federrath , C., & Brunt , C. M. 2011, , 727, L21

  76. [84]

    Ram \' rez , I., Allende Prieto , C., & Lambert , D. L. 2013, , 764, 78

  77. [85]

    2012, , 750, L31

    Robertson , B., & Goldreich , P. 2012, , 750, L31

  78. [86]

    Rosotti, G. P. 2023, New Astronomy Reviews, 96, 101674

  79. [87]

    2024, Science Advances, 10, eado3958

    Scannapieco , E., Pan , L., Buie , II, E., & Br \"u ggen , M. 2024, Science Advances, 10, eado3958

  80. [88]

    A., Cowley , S

    Schekochihin , A. A., Cowley , S. C., Dorland , W., et al. 2009, , 182, 310

  81. [89]

    Schmidt , W., Federrath , C., Hupp , M., Kern , S., & Niemeyer , J. C. 2009, , 494, 127

  82. [90]

    2020, , 132, 104301

    Schneider , N., Simon , R., Guevara , C., et al. 2020, , 132, 104301

  83. [91]

    2020, , 497, 4196

    Seifried , D., Walch , S., Weis , M., et al. 2020, , 497, 4196

  84. [92]

    Squire , J., & Hopkins , P. F. 2017, , 471, 3753

  85. [93]

    M., Ostriker , E

    Stone , J. M., Ostriker , E. C., & Gammie , C. F. 1998, , 508, L99

  86. [94]

    M., Tomida , K., White , C

    Stone , J. M., Tomida , K., White , C. J., & Felker , K. G. 2020, , 249, 4

  87. [95]

    Sur , S., Scannapieco , E., & Ostriker , E. C. 2016, , 818, 28

  88. [96]

    S., & Dopita , M

    Sutherland , R. S., & Dopita , M. A. 1993, , 88, 253

  89. [97]

    W., Leroy , A

    Tamburro , D., Rix , H. W., Leroy , A. K., et al. 2009, , 137, 4424

  90. [98]

    J., & Tan , J

    Tasker , E. J., & Tan , J. C. 2009, , 700, 358

  91. [99]

    2021, Computing in Science Engineering, 23, 10

    Trott, C., Berger-Vergiat, L., Poliakoff, D., et al. 2021, Computing in Science Engineering, 23, 10

  92. [100]

    F., & Blake , G

    van Dishoeck , E. F., & Blake , G. A. 1998, , 36, 317

  93. [101]

    1994, Astrophysical Journal v.423, 423, 681

    Vazquez-Semadeni, E. 1994, Astrophysical Journal v.423, 423, 681

  94. [102]

    1998, , 492, 596

    V \'a zquez-Semadeni , E., Cant \'o , J., & Lizano , S. 1998, , 492, 596

  95. [103]

    2015, , 454, 238

    Walch , S., Girichidis , P., Naab , T., et al. 2015, , 454, 238

  96. [104]

    2019, , 878, 157

    Xu , S., Ji , S., & Lazarian , A. 2019, , 878, 157

  97. [105]

    W., Mantz , A., & Werner , N

    Zhuravleva , I., Allen , S. W., Mantz , A., & Werner , N. 2018, , 865, 53

  98. [106]

    G., & McKee , C

    Zweibel , E. G., & McKee , C. F. 1995, , 439, 779

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.