Numerical approach to compressible shallow-water dynamics of neutron-star spreading layers

Aleksandr Rusakov; Omer Bromberg; Pavel Abolmasov

REVIEW 4 major objections 6 minor 25 references

Numerical approach to compressible shallow-water dynamics of neutron-star spreading layers

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A new spherical hydrodynamics code, SPLASH, solves supersonic neutron-star spreading-layer flows at Mach 5-10 and finds a rigidly rotating two-armed pattern that appears as a quasi-periodic oscillation to inclined observers.

desk verdict Solid open-source code and honest methods tests; the spreading-layer QPO claim outruns the evidence. read the letter →

arxiv 2412.00867 v2 pith:DB2KTV47 submitted 2024-12-01 astro-ph.HE

Aleksandr Rusakov , Pavel Abolmasov , Omer Bromberg This is my paper · ORCID

classification astro-ph.HE

keywords ComputationalastronomyHydrodynamicsAstrophysicalfluiddynamicsStellaraccretionNeutronstarsLow-massX-raybinaryspreadinglayerquasi-periodicoscillations

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

This paper develops SPLASH, a new numerical code that solves the compressible shallow-water equations on a sphere using a multislope second-order MUSCL scheme and an HLLC+ all-speed Riemann solver on an unstructured spherical mesh. The authors' goal is to simulate the accretion spreading layer of a weakly magnetized neutron star in the physically realistic regime, where the flow is supersonic with Mach numbers of order 5-10, a regime that earlier spectral methods could not reach because of Gibbs phenomena at shocks. When they feed a realistic accretion rate onto a 1.5 solar-mass, 10 km neutron star, the equatorial heating triggers a Rayleigh-Taylor instability that mixes the layer in latitude and eventually organizes it into a rigidly rotating two-armed 'tennis-ball' pattern. The paper claims this pattern produces a high-quality quasi-periodic oscillation in the light curve of a high-inclination observer, with a frequency well below the rotation frequency of the matter in the layer, and that the other variability is likely Rossby-wave-like. If correct, the work provides the first numerically grounded, time-dependent model for the variability of accreting neutron stars and a tool for studying any two-dimensional supersonic flow on a sphere.

What carries the argument

The carrying machinery is the SPLASH finite-volume scheme: a multislope second-order MUSCL reconstruction with a hybrid limiter, paired with an HLLC+ all-speed approximate Riemann solver, implemented on a spherical tessellation whose edges are great-circle arcs, so the grid has no polar singularity and can capture shocks at Mach 5-10. The conserved variables are the surface density $\Sigma$, the three Cartesian components of angular-momentum surface density $\mathbf{l}$, and the surface energy density $E$, with an effective two-dimensional adiabatic index $\Gamma = 2 - 1/\gamma$. On the physics side, the argument is carried by the latitudinal entropy gradient produced by accretion heating near the equator: Appendix B derives the Brunt-Väisälä-type criterion for rigid-body rotation, showing that a constant-entropy state is neutral while a constant-density state is convectively unstable, and the paper relates the nonlinear outcome to Rayleigh-Taylor growth with effective gravity $g_{\theta,\rm eff} = \Omega^2 R \sin\theta \cos\theta$. The observable signal is defined by the light-curve proxy $L_c = \sum_i \Pi_i K_i (\mathbf{R}_{\rm obs}\cdot \mathbf{R}_i)$, a pressure-weighted projected area computed for four observer inclinations.

What would settle it

Replace the light-curve proxy with the actual radiative-loss term $\dot{E}_{\rm rad} = -c g_{\rm eff}(1-\beta)/\kappa$ from equation (44), recompute the Lomb-Scargle periodograms from the same snapshots, and check whether the peak at twice the pattern rotation frequency survives; if it vanishes, the claimed quasi-periodic oscillation is an artifact of the diagnostic. A second check is to rerun the accretion case on a finer mesh for several mass-renewal times (≳100 s) and see whether the two-armed pattern and its harmonic persist.

Watch

Extended reading notes

Core claim

The paper makes a two-part claim. First, SPLASH, built on a multislope second-order MUSCL reconstruction and an HLLC+ all-speed Riemann solver on an unstructured spherical tessellation, solves the compressible shallow-water equations with second-order accuracy for Mach numbers up to at least 5-10, as demonstrated by stationary-atmosphere, shock, and split-sphere tests. Second, when applied to constant accretion onto a spherical neutron star, the code shows that equatorial heating triggers a Rayleigh-Taylor instability that mixes the layer in latitude and saturates into a rigidly rotating two-armed 'tennis-ball' pattern. The pattern appears in the computed light curves, defined as the pressure-weighted projected area $L_c = \sum_i \Pi_i K_i (\mathbf{R}_{\rm obs}\cdot \mathbf{R}_i)$, as a high-quality quasi-periodic oscillation for high-inclination observers, at a frequency well below the local matter rotation frequency, with additional variability attributed to Rossby modes. The authors conclude that this is the first simulation of a spreading layer in the realistic parameter range appropriate for a weakly magnetized accreting neutron star.

Load-bearing premise

The load-bearing premise is that the pressure-weighted projected area of the simulated layer, equation (63), behaves like the radiation a real telescope would detect; the paper's timing analysis and QPO interpretation rest entirely on this proxy rather than on the actual radiative loss rate.

Editorial extensions

If this is right

SPLASH can now be used for any two-dimensional hydrodynamic problem on a sphere with Mach numbers up to at least 5-10, including shock-dominated flows, without the polar singularities or Gibbs oscillations that limited earlier spectral approaches.
If the 'tennis-ball' pattern is generic, accreting neutron stars should show a high-quality quasi-periodic oscillation near twice the pattern rotation frequency to high-inclination observers, at a frequency well below the local Keplerian rate.
The predicted hemisphere asymmetry, about 20% difference in mean rotation rates by the end of the run, implies that variability and light-curve properties may depend on which hemisphere faces the observer.
The same simulation framework, with friction and mass-sink terms balanced, should be able to reach a steady state and test whether the pattern and its oscillation modes survive beyond the initial transient accretion phase.

Reading between the lines

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

The QPO claim should be tested by replacing the pressure-weighted projected-area diagnostic with a proper radiative-transfer calculation of escaping flux; if the $2\nu_{\rm cycl}$ peak disappears, the mode is a property of the diagnostic, not of the star.
The topological selection of azimuthal number $m=2$ suggests a testable family of predictions: narrower or tilted accretion bands, or different initial rotation profiles, should excite $m=1$ or $m=3$ patterns with correspondingly different harmonic structure.
If the pattern survives into a steady state, the ratio of the QPO frequency to the neutron star spin frequency becomes a probe of the layer's rotation profile and of the observer's inclination, which could be compared directly with LMXB timing observations.

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.

Desk Editor's Note

read the letter

Bottom line: this is a solid numerical methods paper with an astrophysical application that is more suggestive than conclusive. The code SPLASH is real, public, and tested: conservation to ~1e-14, stable rigid rotation for ten periods, and a Mach-6 shock test on irregular spherical meshes. That part holds up. The multislope MUSCL plus HLLC+ combination on spherical tessellations appears genuinely new for this problem, and the previous spectral approach could not handle shocks. The tennis-ball pattern and the hemispheric asymmetry are new results from the simulation, not fitting artifacts.

The soft spots are concentrated in the astrophysical section. The main run is a single simulation, about one second long, with no resolution study and no control without the 6-degree tilt. The authors are upfront about the lack of steady state and the missing friction in Section 5.1, which earns credit. The weakest link is the light-curve claim. Lc in Eq. (63) is a pressure-weighted projected area, and Section 5.3 concedes it does not necessarily affect the radiative energy loss term. All periodograms and the QPO discussion use Lc, so the paper shows a pattern in the hydrodynamic fields but does not show that a real observer's flux tracks it. The periodogram peaks are interpreted without significance thresholds or error bars, and the v1 abstract's 'high-quality-factor QPO' overclaims; the v2 abstract is more careful. The 6-degree tilt is purposeful, but since the asymmetry is described as spontaneous, a zero-tilt control would clarify how much of the lopsidedness is source geometry.

The citation pattern is reasonable. Prior SL work is cited, and the self-citation to Abolmasov et al. 2020 is appropriate because this directly extends that spectral method. No invented entities or hidden circular fits here.

Who gets value from this: anyone working on numerical shallow-water on spheres, and anyone studying spreading layers or boundary layers in LMXBs. It deserves a serious referee. I would send it to review with the expectation that the astrophysical claims get tightened or the paper is reframed as a methods paper with a preliminary application. The code and tests are worth publishing even if the QPO interpretation fades.

Referee Report

4 major / 6 minor

Summary. The paper develops SPLASH, a new two-dimensional hydrodynamics code that solves compressible shallow-water equations on an arbitrary irregular spherical mesh using a multislope second-order MUSCL scheme with an HLLC+ Riemann solver. After verifying conservation and stability on stationary rotating atmospheres, a shock test, and a split-sphere shear test, the code is applied to an accreting neutron-star spreading layer with sources and sinks for mass, angular momentum, and energy. The simulation develops a convective instability and subsequently a global two-armed 'tennis ball' pattern rotating nearly rigidly, accompanied by long-lived cyclones. Light curves are computed for four observer inclinations via a pressure-weighted projected area diagnostic, and time-resolved Lomb-Scargle periodograms show peaks near harmonics of the mean rotation and cyclone pattern frequencies. The authors conclude that the pattern causes observable flux variations at high inclination and that the code can simulate spreading layers in a realistic parameter range.

Significance. If the numerical method holds up, this is a useful contribution: SPLASH is publicly available, conserves mass and energy to roughly machine precision in the tested stationary case, and demonstrates stable behavior at Mach numbers of order 5-10 on spherical unstructured meshes without polar singularities. The spreading-layer simulation also produces a striking, long-lived two-armed pattern that is physically interesting and worthy of further study. However, the central astrophysical claim—that the pattern produces detectable flux variability for a real observer—rests entirely on a diagnostic, Eq. (63), that the authors themselves state need not track the radiative energy loss. The single simulation also lacks a resolution study, and the paper's own Section 5.1 concedes that the run is far from a realistic steady state and omits surface friction. The observational conclusion is therefore not yet established, even though the code-oriented part of the paper is largely sound.

major comments (4)

[§5.3, Eq. (63)] The 'light curve' Lc is a pressure-weighted projected area, not a radiative flux. The paper explicitly notes that this diagnostic 'outline[s] the most dynamically violent processes in the layer, which do not necessarily affect the radiative energy loss term (equation 44).' Nevertheless, all periodograms in Fig. 13, the QPO discussion, and the Section 7 conclusion that the pattern 'causes flux variations for an observer at a large inclination' are derived from Lc. As written, the paper does not demonstrate that a real observer's flux would follow Lc. I request either a validation of Lc against a flux-based diagnostic (e.g., a light curve constructed from the actual cooling term of Eq. 44, or a simple emission model using the local effective temperature) or a substantial toning-down of the observational claims so that they apply only to the pressure diagnostic.
[§5.1, §5.2, Fig. 10] The spreading-layer result is based on a single simulation with one cubic mesh of 24,576 faces. Section 4 tests the scheme on stationary and shock problems, but there is no resolution study for the full accretion run, so it is unknown whether the two-armed pattern, its rotation frequency, and the cyclone lifetimes are converged or mesh-dependent. A coarser and a finer run (or at least a second mesh type at similar resolution) are needed to support the claim that the pattern is a robust physical outcome rather than a low-resolution artifact.
[§5.1 and §7] The conclusion that this is 'for the first time to simulate the dynamics of an accretion spreading layer in the realistic parameter range appropriate for a real weakly magnetized accreting NS' contradicts the limitations stated in Section 5.1: the simulation covers only about one second, does not reach mass equilibrium, and does not include friction with the NS surface, which the authors identify as 'a crucial constituent for angular momentum conservation and energy release.' The word 'realistic' should either be removed or the claim restricted to the early dynamical phase that is actually simulated.
[§5.3, §6] The identification of the low-frequency peaks near νcycl/2 and νcycl as 'probably being a result of some inertial mode' is not tested. Section 6 lists several possible modes (Rossby, gravity, pressure) but does not compare the observed peak frequencies to the dispersion relations, except by qualitative statements about 'likely' origins. A mode identification would require, for example, checking the azimuthal wavenumber and latitudinal structure of the oscillations, or comparing with the Rossby-wave relation of Eq. (64) at specific latitudes. As it stands, the mode assignments in the abstract and conclusion are speculative.

minor comments (6)

[Title page] The author list contains a typo: 'Pa velAbolmasov' should presumably read 'Pavel Abolmasov.'
[Section 4, footnote 1] The GitHub URL is set in quotes as `https://github.com/TURBOLOSE/SPLASH' with a stray quotation mark; this should be cleaned up.
[Section 2.3, Eq. (21)] The physical meaning of the relation between H, Π, Σ, and geff would be clearer if the authors stated explicitly that Π is the vertically integrated pressure before introducing Eq. (21).
[Section 5.3] The description of the light-curve normalization is incomplete: the units of Lc in Eq. (63) are not stated, and the figure caption for Fig. 13 does not say whether the vertical axis is normalized frequency or physical frequency. Please make the units and normalization explicit.
[Section 2.5, Eq. (35)] The Gaussian accretion source uses σα as the width but the text later refers to σ = 6°; for clarity, use the same symbol in both places.
[Appendix B, Eq. (B18)] The text says the increment is 'always imaginary, meaning convective stability,' but the sign of N2 and the condition for instability in the Boussinesq analysis are not fully discussed. Please clarify whether the sign convention is chosen so that N2 < 0 corresponds to instability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hydrodynamic derivation is self-contained and benchmarked, and the Lc light-curve proxy is a diagnostic limitation rather than a circular reduction.

full rationale

The derivation chain is self-contained. The compressible shallow-water conservation laws (Eqs. 1-3, 10, 14, 26) are derived in Section 2, and the numerical scheme is validated against the analytic rigid-rotation solution of Appendix A (mass/energy conservation to ~1e-14, profiles preserved over 10 rotations), a Mach-6 shock test, and a split-sphere Kelvin-Helmholtz test. The beta/gamma closure (Eqs. 29-32) and the radiation sink (Eq. 44) are adopted from Abolmasov et al. (2020), a published derivation involving one of the present authors; these are independent, externally falsifiable physical inputs, not uniqueness theorems, and they do not encode the simulation outcome. The 'tennis-ball' pattern rotation frequency is measured by tracking a pressure minimum (nu_cycl), and the periodogram peaks are computed from a separately defined diagnostic Lc (Eq. 63); identifying the peak at 2 nu_cycl as the pattern's second harmonic is an interpretation of independently measured quantities, not a fit of the periodogram to nu_cycl. The only notable weakness is that Lc is a pressure-weighted projected area, and the paper itself cautions that it 'do[es] not necessarily affect the radiative energy loss term (equation 44)'; this limits the astrophysical flux prediction but does not feed back into the equations and does not make the result equivalent to an input. No circular step is exhibited.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new particles, forces, or conserved quantities. The tennis-ball pattern and cyclones are emergent flow features, not postulated entities. The free parameters are initial conditions and physical inputs of the simulation; none are fitted to a target observational result. The main unproven inputs are the domain assumptions of the vertically integrated model, the radiation-loss prescription, the light-curve proxy, and the absence of surface friction, all of which are stated by the authors.

free parameters (6)

Initial polar Mach number M0 = 5
Sets the rotation speed of the initial rigidly rotating atmosphere relative to sound speed; chosen for code stability rather than from observations (Section 5.1).
Initial rotation frequency Omega = 300 s^-1
Rotation of the initial atmosphere; selected relatively low to improve code stability (Section 5.1).
Accretion rate Mdot = 1e-8 solar masses per year
Typical Z-source accretion rate used as a physical input (Section 2.1).
Injection azimuthal velocity v_orb = 0.4c
Slightly below the Keplerian value 0.47c; sets the angular momentum input (Section 5.1).
Accretion band tilt and width = alpha_tilt=6 degrees, sigma_alpha=6 degrees
Tilt breaks axial symmetry and width mimics a thin disk; both chosen by hand rather than derived (Section 5.1).
Mass sink time tfall = Mtot/Mdot
Defined from global mass balance so the layer mass stays roughly constant; effectively a free decay time (Section 2.5).

assumptions (6)

domain assumption The flow is vertically integrated and thin (h << R); only tangential velocity components are retained.
Justifies the two-dimensional shallow-water system in Section 2.1 and the neglect of vertical velocities.
domain assumption Vertical structure is polytropic with gamma_v = 4/3, set by radiation transfer with constant opacity; the effective two-dimensional index Gamma depends on the gas-to-total pressure ratio beta.
Underlies the vertically integrated energy and flux expressions (equations 25 through 31).
domain assumption No friction between the spreading layer and the star's crust.
Explicitly stated as a limitation in Sections 5.1 and 7; angular momentum is otherwise removed only via the mass sink.
domain assumption Opacity is Thomson (kappa_T = 0.34 cm^2/g) and radiation escapes at the rate given by equation (44).
Radiation sink prescription from Abolmasov et al. (2020), used in the energy equation.
domain assumption The light-curve proxy Lc = sum Pi_i K_i (R_obs dot R_i) represents what a distant observer sees.
All timing analysis in Section 5.3 uses this proxy, which the authors note does not necessarily affect the radiative energy loss term.
standard math The local linear analysis in Appendix B under the Boussinesq approximation describes the instability.
Used for the convective and Rayleigh-Taylor instability increment estimates, equations (B16) through (B18).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Numerical approach to compressible shallow-water dynamics of neutron-star spreading layers." pith.science (2026). https://pith.science/paper/DB2KTV47

@misc{pith2026241200867,
  author       = {Pith},
  title        = {Pith review of: Numerical approach to compressible shallow-water dynamics of neutron-star spreading layers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DB2KTV47}},
  note         = {Machine review of arXiv:2412.00867}
}

read the original abstract

A weakly magnetized neutron star (NS) undergoing disk accretion should release about a half of its power in a compact region known as the accretion boundary layer. Latitudinal spread of the accreted matter and efficient radiative cooling justify the approach to this flow as a two-dimensional spreading layer (SL) on the surface of the star. Numerical simulations of SLs are challenging because of the curved geometry and supersonic nature of the problem. We develop a new two-dimensional hydrodynamics code that uses the multislope second-order MUSCL scheme in combination with an HLLC+ Riemann solver on an arbitrary irregular mesh on a spherical surface. The code is suitable and accurate for Mach numbers at least up to 5-10. Adding sinks and sources to the conserved variables, we simulate constant-rate accretion onto a spherical NS. During the early stages of accretion, heating in the equatorial region triggers convective instability that causes rapid mixing in latitudinal direction. One of the outcomes of the instability is the development of a two-armed `tennis ball' pattern rotating as a rigid body. From the point of view of a high-inclination observer, its contribution to the light curve is seen as a high-quality-factor quasi-periodic oscillation mode with a frequency considerably smaller than the rotation frequency of the matter in the SL. Other variability modes seen in the simulated light curves are probably associated with low-azimuthal-number Rossby waves.

Figures

Figures reproduced from arXiv: 2412.00867 by the authors.

**Figure 2.** Geometric reconstruction for 2D configuration. In order to apply the algorithm to a spherical mesh, several adaptations have to be made. We can treat meshes either like a polyhedra or as spherical tessellations, as shown [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 3.** Difference between a 2D mesh, a spherical polyhedron mesh and a spherical tessellation. Additionally, we replace normals n with −(n×R) in angular momentum fixes that were introduced in HLLC+. As for the limiter, we used the hybrid limiter introduced in Touze et al. (2015). It is proven to be stable for multislope MUSCL reconstruction, and it was the one that worked best in most of the tests. 3.4. Time-step stability… view at source ↗

Figures from the paper (10 more)

**Figure 4.** Figure 4: Relative error in conservation of total mass and total energy [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

**Figure 5.** Figure 5: Conservation of Σ and Π profiles after 10 full rotations on a hexagonal mesh with 10242 faces [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

**Figure 6.** Figure 6: Conservation of Σ and Π profiles after 10 full rotations on a triangular mesh with 20480 faces [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

**Figure 7.** Figure 7: Conservation of Σ and Π profiles after 10 full rotations on a cubic mesh with 24576 faces [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

**Figure 8.** Figure 8: Pressure maps in the shock test. as oscillations and instabilities that are expected to emerge during the formation of an SL or a rapid increase in mass accretion rate. Results of the simulations are stored as snapshots of all the primitives in all the cells saved sev…

**Figure 9.** Figure 9: Vorticity plots for the split-sphere test. The merge of multiple smaller vortices into a bigger one can be observed. considerably for the northern and southern hemispheres, as demonstrated in [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

**Figure 10.** Figure 10: Density maps of the accretion simulation. Six snapshots: initial state, instability start, instability evolution, and the final state. Full video of simulation is available on YouTube: https://youtu.be/BVzGHAnDtwA [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]

**Figure 11.** Figure 11: Final state of accretion simulation with streamlines. First plot contains streamlines with rotation, the second is without average rotation [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]

**Figure 12.** Figure 12: Average rotational frequency of northern and southern hemispheres. Starting from t ∼ 0.07s, there is a measurable difference in the rotation velocities and stream patterns between the Northern and the Southern hemispheres, reaching about 20% by the end of the simulat…

**Figure 13.** Figure 13: Dynamic power spectra of the accretion simulation for different observers. Overplotted are the z-component of the average frequency, its first harmonic, cyclone frequency with its first harmonic and the frequency sum. A. ANALYTIC SOLUTIONS FOR RIGID-BODY ROTATION Hydr…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 8 canonical work pages

[1]

2020, A&A, 638, A142, doi: 10.1051/0004-6361/201936958

Abolmasov, P., N¨attil¨a, J., & Poutanen, J. 2020, A&A, 638, A142, doi: 10.1051/0004-6361/201936958

work page doi:10.1051/0004-6361/201936958 2020
[2]

2021, A&A, 647, A45, doi: 10.1051/0004-6361/202039485

Abolmasov, P., & Poutanen, J. 2021, A&A, 647, A45, doi: 10.1051/0004-6361/202039485

work page doi:10.1051/0004-6361/202039485 2021
[3]

2023, in Handbook of X-ray and Gamma-ray Astrophysics, 120, doi: 10.1007/978-981-16-4544-0 94-1

Bahramian, A., & Degenaar, N. 2023, in Handbook of X-ray and Gamma-ray Astrophysics, 120, doi: 10.1007/978-981-16-4544-0 94-1

doi:10.1007/978-981-16-4544-0 2023
[4]

2018, MNRAS, 473, 2771, doi: 10.1093/mnras/stx2508

Bransgrove, A., Levin, Y ., & Beloborodov, A. 2018, MNRAS, 473, 2771, doi: 10.1093/mnras/stx2508

doi:10.1093/mnras/stx2508 2018
[5]

1961, Hydrodynamic and hydromagnetic stability

Chandrasekhar, S. 1961, Hydrodynamic and hydromagnetic stability

1961
[6]

2020, SIAM Journal on Scientific Computing, 42, B921, doi: 10.1137/18M119032X

Chen, S., Lin, B., Li, Y ., & Yan, C. 2020, SIAM Journal on Scientific Computing, 42, B921, doi: 10.1137/18M119032X

work page doi:10.1137/18m119032x 2020
[7]

2003, A&A, 410, 217, doi: 10.1051/0004-6361:20031141

Gilfanov, M., Revnivtsev, M., & Molkov, S. 2003, A&A, 410, 217, doi: 10.1051/0004-6361:20031141

doi:10.1051/0004-6361:20031141 2003
[8]

1997, SIAM Review, 39, 644, doi: 10.1137/S0036144596301390

Gottlieb, D., & Shu, C.-W. 1997, SIAM Review, 39, 644, doi: 10.1137/S0036144596301390

doi:10.1137/s0036144596301390 1997

Show all 25 references

[9]

1989, A&A, 225, 79

Hasinger, G., & van der Klis, M. 1989, A&A, 225, 79

1989
[10]

R., & Motta, S

Ingram, A. R., & Motta, S. E. 2019, NewAR, 85, 101524, doi: 10.1016/j.newar.2020.101524

2019
[11]

A., & Sunyaev, R

Inogamov, N. A., & Sunyaev, R. A. 1999, Astronomy Letters, 25, 269, doi: 10.48550/arXiv.astro-ph/9904333 —. 2010, Astronomy Letters, 36, 848, doi: 10.1134/S1063773710120029

1999 doi arXiv
[12]

J., & Williamson, D

Jakob-Chien, R., Hack, J. J., & Williamson, D. L. 1995, Journal of Computational Physics, 119, 164, doi: 10.1006/jcph.1995.1125

1995
[13]

Kluzniak, W., Michelson, P., & Wagoner, R. V . 1990, ApJ, 358, 538, doi: 10.1086/169006

1990 doi
[14]

D., & Lifshitz, E

Landau, L. D., & Lifshitz, E. M. 1987, Fluid Mechanics

1987
[15]

Lomb, N. R. 1976, Ap&SS, 39, 447, doi: 10.1007/BF00648343 N¨attil¨a, J., Cho, J. Y . K., Skinner, J. W., Most, E. R., & Ripperda, B. 2024, ApJ, 971, 37, doi: 10.3847/1538-4357/ad54c2

1976 doi
[16]

Papaloizou, J. C. B., & Stanley, G. Q. G. 1986, MNRAS, 220, 593, doi: 10.1093/mnras/220.3.593

1986 doi
[17]

Payne, D. J. B., & Melatos, A. 2004, MNRAS, 351, 569, doi: 10.1111/j.1365-2966.2004.07798.x

2004
[18]

Scargle, J. D. 1982, ApJ, 263, 835, doi: 10.1086/160554 21

1982 doi
[19]

I., & Sunyaev, R

Shakura, N. I., & Sunyaev, R. A. 1973, A&A, 24, 337 —. 1988, Advances in Space Research, 8, 135, doi: 10.1016/0273-1177(88)90396-1

1973 doi
[20]

2020, Acta Numerica, 29, 701–762, doi: 10.1017/S0962492920000057

Shu, C.-W. 2020, Acta Numerica, 29, 701–762, doi: 10.1017/S0962492920000057

2020 doi
[21]

2021, Generating Meshes of a Sphere

Sieger, D. 2021, Generating Meshes of a Sphere. https: //danielsieger.com/blog/2021/03/27/generating-spheres.html

2021
[22]

2006, MNRAS, 369, 2036, doi: 10.1111/j.1365-2966.2006.10454.x

Suleimanov, V ., & Poutanen, J. 2006, MNRAS, 369, 2036, doi: 10.1111/j.1365-2966.2006.10454.x

2006
[23]

L., Murrone, A., & Guillard, H

Touze, C. L., Murrone, A., & Guillard, H. 2015, Journal of Computational Physics, 284, 389, doi: 10.1016/j.jcp.2014.12.032

2015 doi
[24]

2024, Galaxies, 12, 43, doi: 10.3390/galaxies12040043 van der Klis, M

Ursini, F., Gnarini, A., Capitanio, F., et al. 2024, Galaxies, 12, 43, doi: 10.3390/galaxies12040043 van der Klis, M. 2000, ARA&A, 38, 717, doi: 10.1146/annurev.astro.38.1.717 —. 2001, ApJ, 561, 943, doi: 10.1086/323378 van Leer, B. 1979, Journal of Computational Physics, 32, ...

2024 doi
[25]

L., Andersson, N., Beyer, H., & Schutz, B

Watts, A. L., Andersson, N., Beyer, H., & Schutz, B. F. 2003, MNRAS, 342, 1156, doi: 10.1046/j.1365-8711.2003.06612.x

2003

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

2020, A&A, 638, A142, doi: 10.1051/0004-6361/201936958

Abolmasov, P., N¨attil¨a, J., & Poutanen, J. 2020, A&A, 638, A142, doi: 10.1051/0004-6361/201936958

work page doi:10.1051/0004-6361/201936958 2020

[2] [2]

2021, A&A, 647, A45, doi: 10.1051/0004-6361/202039485

Abolmasov, P., & Poutanen, J. 2021, A&A, 647, A45, doi: 10.1051/0004-6361/202039485

work page doi:10.1051/0004-6361/202039485 2021

[3] [3]

2023, in Handbook of X-ray and Gamma-ray Astrophysics, 120, doi: 10.1007/978-981-16-4544-0 94-1

Bahramian, A., & Degenaar, N. 2023, in Handbook of X-ray and Gamma-ray Astrophysics, 120, doi: 10.1007/978-981-16-4544-0 94-1

doi:10.1007/978-981-16-4544-0 2023

[4] [4]

2018, MNRAS, 473, 2771, doi: 10.1093/mnras/stx2508

Bransgrove, A., Levin, Y ., & Beloborodov, A. 2018, MNRAS, 473, 2771, doi: 10.1093/mnras/stx2508

doi:10.1093/mnras/stx2508 2018

[5] [5]

1961, Hydrodynamic and hydromagnetic stability

Chandrasekhar, S. 1961, Hydrodynamic and hydromagnetic stability

1961

[6] [6]

2020, SIAM Journal on Scientific Computing, 42, B921, doi: 10.1137/18M119032X

Chen, S., Lin, B., Li, Y ., & Yan, C. 2020, SIAM Journal on Scientific Computing, 42, B921, doi: 10.1137/18M119032X

work page doi:10.1137/18m119032x 2020

[7] [7]

2003, A&A, 410, 217, doi: 10.1051/0004-6361:20031141

Gilfanov, M., Revnivtsev, M., & Molkov, S. 2003, A&A, 410, 217, doi: 10.1051/0004-6361:20031141

doi:10.1051/0004-6361:20031141 2003

[8] [8]

1997, SIAM Review, 39, 644, doi: 10.1137/S0036144596301390

Gottlieb, D., & Shu, C.-W. 1997, SIAM Review, 39, 644, doi: 10.1137/S0036144596301390

doi:10.1137/s0036144596301390 1997