Pith. sign in

REVIEW 4 major objections 4 minor 4 cited by

Bow-shock radio emission from TDE outflows peaks at 1-20 GHz, rises steeper than t^4, and leaves a spectral fingerprint separable from the standard forward-shock component.

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 · deepseek-v4-flash

2026-08-04 07:36 UTC pith:QBRQ3JTK

load-bearing objection First numerical bow-shock TDE radio study with credible qualitative diagnostics, but the headline scalings rest on imposed microphysics and need to be demonstrated before the paper's quantitative claims can be trusted. the 4 major comments →

arxiv 2510.25033 v2 pith:QBRQ3JTK submitted 2025-10-28 astro-ph.HE astro-ph.GA

Numerical Studies on the Radio Afterglows in TDE: Bow Shock

Guobin Mou , Xinwen Shu This is my paper
classification astro-ph.HE astro-ph.GA
keywords Tidal disruption eventsRadio afterglowsBow shockForward shockSynchrotron radiationRelativistic electronsCircumnuclear mediumRadiative transfer
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.

This paper tries to establish that the radio afterglow of a tidal disruption event can be generated by a bow shock, where the outflow collides with a dense cloud near the black hole, and that this bow-shock component is observationally distinguishable from the standard forward-shock emission. Using two-fluid simulations that follow relativistic electrons and compute synchrotron spectra via radiative transfer, the authors find the bow-shock emission peaks at higher frequencies (roughly 1-20 GHz), rises much more steeply than the forward shock (steeper than t^4 in the explored parameter space), declines rapidly, and fluctuates when the outflow fluctuates. When both shocks operate, the composite spectrum deviates from a single-zone model and can show double-peaked or flat-top features. The authors also show that applying the standard equipartition method to bow-shock emission can misestimate the shock radius by a factor of roughly 0.3-3 and the nonthermal energy by 0.03-2. If the picture is right, the shape and variability of TDE radio afterglows become a tool to find dense circumnuclear gas around otherwise quiescent supermassive black holes, without relying on dust.

Core claim

The central claim is that a bow shock wraps around the windward side of a dense cloud hit by a TDE outflow, and its radio emission forms a compact, concentrated nonthermal region whose synchrotron spectrum is self-absorbed at a higher frequency than the forward-shock spectrum. The simulations show peak frequencies of roughly 1-20 GHz that stay nearly constant while the bow shock lasts, then drop to GHz/sub-GHz as the shocked material expands and becomes transparent. The flux rises sharply when the outflow first hits the cloud, stays fairly stable for the bow-shock lifetime, responds to outflow-density fluctuations on month timescales, and declines rapidly when the interaction ends. The autho

What carries the argument

The central machinery is a pair of fluid components — thermal gas and relativistic electrons (CRe) — evolved together in a hydrodynamics simulation, with the electrons injected into the fourth mesh zone behind the shock (where post-shock parameters have settled) and then advected with the downstream flow without diffusion. The magnetic field is not simulated dynamically; in every cell its energy density is set to a fixed fraction of the CRe energy density, (8π)^-1 B^2/e_2 = ε_B/ε_e with ε_e=0.03 and ε_B=0.10. The resulting nonthermal distribution, including a jellyfish-like tail trailing the cloud, is interpolated into 3D and used in radiative-transfer calculations to produce synthetic spect

Load-bearing premise

The result rests on treating the relativistic electrons as injected immediately behind the shock and then frozen to the downstream flow with no diffusion, while the magnetic field energy in each cell is kept in a fixed ratio to the electron energy; if electrons spread out or the field is not locally tied to the electrons, the concentrated emission region that produces the high peak frequency, steep rise, and rapid fluctuations would be smeared out.

What would settle it

Simultaneous low-frequency (around 1 GHz) and high-frequency (10-20 GHz) monitoring of a TDE afterglow over several years: the bow-shock model predicts the high-frequency flux to turn on sharply, show month-scale fluctuations, and decline quickly, with the spectral peak above roughly 1 GHz while the bow shock is active. A light curve that instead rises slowly and peaks at frequencies that drift downward with time, as the forward-shock model predicts, would falsify the bow-shock interpretation for that source. VLBI is a second falsifier: the bow shock's bright spot should stay stationary while

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

If this is right

  • Radio afterglows of TDEs with a sharp delayed rise, a plateau, a rapid decline, and month-scale flux fluctuations are candidates for bow-shock emission rather than forward-shock emission.
  • Multi-frequency spectra that show double-peaked or flat-top features can be decomposed into a high-frequency bow-shock component and a low-frequency forward-shock component, yielding constraints on cloud distance, cloud size, outflow energy, and circumnuclear density.
  • Equipartition-based estimates of shock radius and nonthermal energy from bow-shock-dominated radio data can be off by a factor of roughly 3 in radius and a factor of 0.03-2 in energy, so single-zone interpretations of TDE radio afterglows should be treated with caution.
  • VLBI imaging can separate the two scenarios: the forward shock expands continuously, while bow-shock bright spots (edge-on) or a bright ring (face-on) remain stationary.
  • Because the radio emission does not depend on dust, the bow-shock signal offers a broader way than infrared echoes to search for sub-parsec dense structures around quiescent supermassive black holes.

Where Pith is reading between the lines

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

  • If the bow-shock picture is correct, the current TDE radio sample may be biased toward dense-cloud environments: events with steeply rising, high-frequency radio flares would preferentially select TDEs with clouds close to the black hole, so inferences about average circumnuclear density from radio-detected TDEs should account for this selection.
  • The same two-component decomposition could be applied to existing archival multi-frequency light curves; finding hidden double-peaked SEDs would be a cost-free test of the scenario.
  • The predictions depend sensitively on the assumptions that relativistic electrons do not diffuse and that the magnetic field tracks the electron energy locally; a follow-up simulation with cosmic-ray diffusion or an evolved magnetic field would show how much of the 1-20 GHz peak and the steep rise survives.
  • The mechanism is generic: any unsteady astrophysical outflow running into a dense clump should produce a similarly high-frequency, fast-variable radio component, so the diagnostic may extend beyond TDEs to AGN outflows and stellar-wind collisions.

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

4 major / 4 minor

Summary. The paper presents 2.5-dimensional two-fluid hydrodynamical simulations (ZEUS-3D) of a TDE outflow striking dense circumnuclear clouds, with a second fluid representing relativistic electron (CRe) energy density. The CRe are injected at a fixed mesh zone behind shocks, advected without diffusion, and the magnetic field in each cell is tied to the CRe energy density by a constant ratio. Synthetic synchrotron spectra and light curves are computed for bow-shock (BS) and combined BS+forward-shock (FS) configurations, for spherical and toroidal clouds at 0.03 pc and 0.2 pc. The central claim is that BS radio emission is more compact, peaks at higher frequency, rises and declines more steeply, responds to outflow fluctuations, and produces double-peaked or broadened spectra when combined with FS emission; the equipartition method is also tested against the simulated BS. The abstract further claims a peak frequency of ~1-20 GHz and a flux rise steeper than t^4.

Significance. If the simulations capture the essential physics, the paper offers a genuinely useful diagnostic: a stationary, high-frequency radio component from a bow shock would indicate dense circumnuclear gas around otherwise quiescent SMBHs and could explain delayed, fluctuating, or spectrally complex TDE radio afterglows. The parameter study is systematic, the radiative-transfer post-processing is transparent, and the synthetic radio images provide a falsifiable VLBI prediction. The comparison with the equipartition method is also valuable for observers. The paper is a forward model with openly stated microphysical choices, so circularity is not a concern. However, the headline quantitative claims are not yet backed by explicit fits or robustness tests, and the most distinctive prediction—the high self-absorbed peak frequency—is controlled by modeling assumptions that are not stress-tested.

major comments (4)
  1. [Abstract; §4.1, Figs. 2-3] The quantitative claims in the abstract—peak frequency "typically ~1-20 GHz" and flux "rises more steeply than t^4"—are not demonstrated in the body. Fig. 3 shows light curves at 0.88, 6, and 15 GHz but includes no t^4 reference curve or fitted power-law index; Fig. 2 shows spectra at discrete epochs but no fitted ν_p(t) evolution. Since these are the paper's headline distinctions between BS and FS, please add explicit scaling fits, reference curves, or best-fit exponents, and state over which time interval the t^4 claim holds. Alternatively, temper the abstract to the qualitative statements actually shown.
  2. [§3 and §2.3] The distinguishing 1-20 GHz peak is set by the spatial concentration of nonthermal plasma at the BS head. That concentration is imposed by two modeling choices: CRe are injected at a fixed "4th mesh" behind the shock with no diffusion, and B^2 is set proportional to e2 in every cell. The paper acknowledges the omission of diffusion (§3) but does not quantify its impact. A modest diffusive precursor, a finite injection region, or independent decay of B downstream would enlarge the effective emitting volume, reduce self-absorption, and likely shift ν_p downward, potentially eroding the BS/FS contrast. Please add sensitivity tests—e.g., varying the injection width, adding a simple diffusion term, or letting B decay on a physical length scale—and show how ν_p and the light-curve shape change.
  3. [§3] The statement "The high resolution adopted here is sufficient to ensure the results are convergent" is asserted without evidence. Because CRe are injected at the numerically resolved 4th mesh zone, the effective injection width is resolution-dependent, which directly affects the compactness of the emitting region and hence self-absorption. A convergence test with at least two additional resolutions for one representative model should be shown before the quantitative peak-frequency claims can be accepted.
  4. [§4.4, Table 2] The equipartition test compares the inferred radius R_eq to the cloud radius R_c, described as the "actual" emission-region radius. For the toroidal and jellyfish morphologies shown in Fig. 5, the emitting region is not simply the cloud radius. The reported factor-of-3 radius accuracy is therefore not well defined. Please define a simulation-measured emission radius (e.g., the effective spherical radius enclosing a given fraction of nonthermal energy) and recompute the ratios, or state clearly that R_c is only a proxy.
minor comments (4)
  1. [Abstract] The arXiv abstract states "peaks at higher frequencies, typically ~1-20 GHz" and "rises more steeply than t^4," while the full-text abstract only says "higher peak frequency" and "much steeper rise." The two versions should be aligned, especially since the stronger quantitative version is one of the main claims.
  2. [Fig. 7 and Table 1] In the upper panel of Fig. 7, the legend lists "fF 0.55 yr" twice; one entry should likely be another model or a different epoch. Also, model cBsv in Table 1 is referred to as "CBsv" in §4.3; use a consistent case.
  3. [§2.3] Equation (3) appears to have a typographical issue: the denominator should likely be (1/2) ρ_i v_s^3 (1 - C^{-2}), with a closing parenthesis after C^{-2}, and the text immediately following uses v_d and e_d without defining them in that context. Please clarify.
  4. [References] The Steinberg & Stone reference lacks a year and journal/volume (arXiv:2206.10641), and the Mou (2025) Paper I reference is cited for many technical details. Since this manuscript is a companion paper, please ensure the reader can reproduce the radiative transfer and CRe injection setup without access to Paper I; at minimum, restate the electron power-law index p and the γ_min/γ_max used for the synchrotron emissivity.

Circularity Check

0 steps flagged

No significant circularity: the BS/FS spectral contrast is a forward prediction from a stated two-fluid simulation, not a renamed fit or a self-citation-forced result.

full rationale

The paper's central claims—BS peak frequency around 1–20 GHz, steep rise and rapid decline, flux fluctuations, and multi-component spectra—are obtained by solving the hydrodynamic equations (4)–(7) for a specified outflow and cloud, evolving a second relativistic-electron fluid, and then performing radiative transfer on the simulated CRe/B distribution. No observed TDE radio data are fitted in this paper; the microphysical inputs (ε_e=0.03, ε_B=0.10, injection at the fourth mesh zone, no CRe diffusion, cell-by-cell B/e2 proportionality) are stated hypotheses, not quantities defined in terms of the target spectra. The spatial concentration of nonthermal plasma at the BS head is partly imposed by the injection scheme, but this is an explicit modeling assumption and a robustness concern, not a derivation that reduces the output to its input. The FS comparison is supported both by FS-only simulations in the present work (models eF and fF in Table 1 and Figs. 6–7) and by Paper I; the latter is a self-citation but not a circular reduction, and it is not invoked as a uniqueness theorem. Section 4.4's adjustment of k_bow from 10 to 1 in the authors' earlier analytical model is an honest post-hoc recalibration, not a fitted parameter disguised as a prediction. The paper itself flags the relevant limitations (Sec. 2.2: 'the setup of the outflow parameters is hypothetical' and 'the results of the BS scenario are preliminary'; Sec. 3: 'we did not include physical diffusion process for the CRe'). These are physical/correctness caveats—possibly weakening the robustness of the claimed peak-frequency diagnostics—but they do not make the derivation circular. Overall, no circular step is identifiable from the text.

Axiom & Free-Parameter Ledger

8 free parameters · 8 axioms · 0 invented entities

The central spectral predictions rest on a chain of chosen microphysical and environmental inputs: the CRe acceleration efficiency and magnetic energy fraction, the outflow launch parameters, cloud density/distance/size, and the Galactic-center-like CNM profile. Most are explicitly hypothetical ('the setup of the outflow parameters is hypothetical', Sec 2.2), so the absolute flux levels are not predicted from first principles; only the qualitative contrasts between BS and FS are robust to these choices.

free parameters (8)
  • epsilon_e (CRe acceleration efficiency) = 0.03
    Adopted in Sec 2.3; sets the energy available for synchrotron-emitting electrons and therefore normalizes all radio fluxes.
  • epsilon_B (magnetic energy fraction) = 0.10
    Adopted in Sec 2.3; sets B via (8π)^{-1}B^2/e2 = epsilon_B/epsilon_e, directly setting synchrotron luminosity and peak frequency.
  • Outflow mass rate Mdot0 = 0.02 or 0.1 Msun/yr
    Hypothetical values (Sec 2.2); set outflow kinetic power (3.4e50 or 1.7e51 erg) and BS ram pressure.
  • Outflow velocity profile v0=0.2c, alpha=2/3 = v0=0.2c, alpha=2/3, Delta t=1 yr
    Chosen velocity decay law v_out = v0(1 - alpha t/Delta t); determines BS duration and light-curve evolution.
  • Cloud density rho_c = 1e6 m_H cm^-3 (dc=0.2 pc); 3e7 m_H cm^-3 (dc=0.03 pc)
    Chosen to keep clouds stable; sets self-absorption and peak frequency.
  • Cloud distance dc and radius Rc = dc=0.03 or 0.2 pc; Rc=0.003-0.04 pc per Table 1
    Key geometric parameters; distance sets ram pressure (~dc^-2), radius sets emission region size and covering factor.
  • CNM density normalization rho0 and index n = rho0=3 m_H cm^-3, n=1.0
    Assumed Galactic-center-like profile (Sec 2.1); sets the FS baseline that BS is compared against.
  • k_bow (adiabatic cooling parameter) = adjusted from 10 to 1
    Sec 4.4: adjusted retroactively so the prior analytical BS model matches the spherical-cloud simulation; not used in the simulation itself but shows the analytical cross-check is partially tuned.
axioms (8)
  • standard math ZEUS-3D hydrodynamic equations and solver correctly model the outflow-cloud interaction in 2.5D spherical coordinates.
    Used in Sec 3; standard but unverified here.
  • domain assumption Shocks efficiently accelerate electrons and amplify magnetic fields to a few percent of ram pressure via resonant streaming instability.
    Invoked in Sec 2.3 citing Völk et al. 2005; not simulated.
  • ad hoc to paper B-field energy density is proportional to CRe energy density in every cell: (8π)^{-1}B^2/e2 = epsilon_B/epsilon_e.
    Sec 2.3; this local coupling is assumed, not derived from MHD.
  • ad hoc to paper CRe are injected at the 4th mesh zone behind the shock and do not diffuse; they are advected with the downstream gas.
    Sec 3; 'We did not include physical diffusion process for the CRe.' This shapes the spatial distribution of nonthermal energy and hence spectra.
  • domain assumption The circumnuclear medium follows a Galactic-center-like power-law profile rho = rho0 r^-1.
    Sec 2.1; sets the FS comparison baseline.
  • domain assumption The TDE outflow parameterization (velocity decay, constant mass rate, 1-yr launch window) is representative of real TDE outflows.
    Sec 2.2 states 'the setup of the outflow parameters is hypothetical'.
  • domain assumption Clouds remain stable because rho_c >> outflow density.
    Sec 2.1; cloud disruption is not modeled.
  • domain assumption The 3D interpolation and radiative transfer procedures from Paper I are correct.
    Sec 3; details are not reproduced in this paper.

pith-pipeline@v1.3.0-alltime-deepseek · 12413 in / 15816 out tokens · 143452 ms · 2026-08-04T07:36:35.259819+00:00 · methodology

0 comments
read the original abstract

The origin of radio afterglows or delayed radio flares in tidal disruption events (TDEs) is not fully understood. They could be generated either by a forward shock propagating into diffuse circumnuclear medium (CNM), or a bow shock around a dense cloud, each of which is fundamentally different. To elucidate the distinctions between these two scenarios, we conducted two-fluid simulations incorporating relativistic electrons to investigate the spatial evolution of these electrons after being accelerated by shock. Based on their spatial distribution, we performed radiative transfer calculations to obtain the synchrotron spectra. In Paper I (Mou 2025), we reported the results for the forward shock scenario; in this article, we focus on the bow shock scenario. Compared to that from the forward shock whose peak frequency typically lies around GHz and decreases with time, the radio emission from the bow shock peaks at higher frequencies, typically $\sim$1-20 GHz, and its flux rises more steeply than $t^4$ across our explored parameter space. The radio flux from the bow shock also responds to fluctuations in the outflow. The combined effects of the bow shock and forward shock substantially alter radio spectra, causing significant deviations from the single-zone emission model, and in some cases producing multi-component feature in spectra. This study highlights the importance of the bow shock, and inspires a novel approach for probing dense gas on sub-parsec scales in galactic nuclei by decomposing the bow shock radio spectrum to reveal the conditions of circumnuclear dense gas.

Figures

Figures reproduced from arXiv: 2510.25033 by Guobin Mou, Xinwen Shu.

Figure 1
Figure 1. Figure 1: BS formed by outflow impacting a toroidal (model ABt) and spherical (model BBs) cloud. The BS of the toroidal cloud is 2D-shock in nature and appears more extended. et al. 2022). Accordingly, the mass of the outflow is expected to reach 10−2 − 10−1M⊙ for disrupting a sun-like star. We parameterize the injected outflow with its velocity vout(t) and mass outflow rate M˙ out(t), both of which should be time￾d… view at source ↗
Figure 2
Figure 2. Figure 2: The synthetic radio spectra along the polar direction (Pol) and the equatorial direction (Eqt) for model ABt and BBs. 10 1 10 0 Time/yr 10 2 10 1 10 0 10 1 F / mJy (d L = 1 0 0 M p c) aBt: 0.88 GHz aBt: 6.0 GHz aBt: 15 GHz bBs: 0.88 GHz bBs: 6.0 GHz bBs: 15 GHz cBsv: 0.88 GHz cBsv: 6.0 GHz cBsv: 15 GHz [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Radio flux along the polar/Z direction for pure BS models (aBt, bBs, cBsv). decreases approximately as d −2 c . For a given high frequency of νi > νp, the monochromatic luminosity Lνi ∝ B (p+1)/2 (Lon￾gair 2011), suggesting that the Lνi approximately decreases as d −(p+1)/2 c (assuming an equal total CRe energy). This indicates clouds located in the inner region are more likely to dominate the radio emissi… view at source ↗
Figure 4
Figure 4. Figure 4: For a cloud located far away (model DBt), the BS apparently un￾dergoes 3 distinct stages: sweeping up post-shock CNM (3.6 yr), post￾shock outflow (4.0 yr), and unshocked outflow (5.5 yr). In the snapshot at t = 3.6 yr, region A, B, C and D mark the CNM, post-shock CNM (by FS), post-shock outflow (by reverse shock) and unshocked outflow, respectively. The synthetic spectra shown are for these 3 epochs along… view at source ↗
Figure 6
Figure 6. Figure 6: Synthetic radio spectra of model aBFt, and its corresponding no￾cloud case – model eF. These results demonstrate the significant roles of cloud and the BS in the overall radio emission. by several tens of times at certain times ( [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 5
Figure 5. Figure 5: Simulation including both the BS and the FS (model aBFt). The left and right panel share the same colorbars. spite the differences, we note that the analytical results for νp and Lνp (Equation 12 and 13 that paper) can match the current synthetic spectra (model bBs in [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
Figure 8
Figure 8. Figure 8: A scheme for exploring the physics of the outflow and CNM environment by multi-frequency radio monitoring. 3 yr 4 yr 6 yr Edge-on (0.9 GHz) Face-on (0.9 GHz) Edge-on (15 GHz) Face-on (15 GHz) [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
Figure 7
Figure 7. Figure 7: Synthetic radio spectra for close cloud cases (upper panel) and distant cloud cases (lower panel). of the circumnuclear environment when both shocks coexist. Specifically, the two- or multi-component radio spectrum (if confirmed) can be analyzed separately, with the high-frequency component corresponding to the BS and the low-frequency com￾ponent to the FS. Combining information from both compo￾nents, it i… view at source ↗
Figure 9
Figure 9. Figure 9: Synthetic radio maps in 2 × 2 mas2 for model eBFt (a toroidal cloud located at dc = 0.2 pc). The axes are labeled in milliarcseconds (dL = 100 Mpc and dA = 96 Mpc). The three epochs correspond to the phases before the BS emerges (3 yr when only the FS exists), the early stage (4 yr) and the late stage (6 yr) of the outflow-cloud interaction. 1. The radio emission from the BS arises from quite efficient sho… view at source ↗

discussion (0)

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

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Radio and X-ray flux rebrightening six years after outburst in a partially-obscured extreme changing-look AGN

    astro-ph.HE 2026-07 conditional novelty 6.5

    SDSS J1548+2208 is a rare partially-obscured extreme CLAGN that launched a radio outflow, with late-time rebrightening explained by interaction with dense circumnuclear clouds.

  2. Hydrodynamical simulation of wind production from hot accretion flows in tidal disruption events

    astro-ph.HE 2026-05 unverdicted novelty 6.0

    Larger black holes in TDEs accrete more material and launch faster winds, with unbound mildly relativistic winds for higher viscosity parameters and bound convective outflows for lower ones.

  3. VLBI Diagnostics of Off-axis Jets in Radio Flares of Tidal Disruption Events

    astro-ph.HE 2025-12 conditional novelty 6.0

    Synthetic VLBI images show that an off-axis jet in AT20118hyz would display superluminal centroid motion and a non-monotonic image aspect ratio, cleanly separating it from a delayed sub-relativistic outflow.

  4. Simulations of interaction between outflow and surrounding broken power-law circumnuclear medium: implications for different radio light curves of TDEs

    astro-ph.HE 2026-06 unverdicted novelty 5.0

    3D hydro simulations show that TDE outflow interactions with a broken power-law CNM can reproduce the range of observed radio light curves via early flares inside the Bondi radius and possible late rebrightenings outside it.

Reference graph

Works this paper leans on

36 extracted references · 3 linked inside Pith · cited by 4 Pith papers

  1. [1]

    D., van Velzen, S., Horesh, A., & Zauderer, B

    Alexander, K. D., van Velzen, S., Horesh, A., & Zauderer, B. A. 2020, Space Sci. Rev., 216, 81

    2020

  2. [2]

    2013, , 772, 78

    Barniol Duran, R., Nakar, E., & Piran, T. 2013, , 772, 78

    2013

  3. [3]

    Bell, A. R. &Lucek, S. G. 2001, , 321, 433

    2001

  4. [4]

    S., Giannios, D., Metzger, B

    Bloom, J. S., Giannios, D., Metzger, B. D., et al. 2011, Science, 333, 203

    2011

  5. [5]

    M., & Lodato, G

    Bonnerot, C., Rossi, E. M., & Lodato, G. 2016, , 458, 3324

    2016

  6. [6]

    N., Kennea, J

    Burrows, D. N., Kennea, J. A., Ghisellini, G., et al., 2011, , 476, 421

    2011

  7. [7]

    2014, , 783, 91

    Caprioli, D., & Spitkovsky, A. 2014, , 783, 91

    2014

  8. [8]

    D., et al

    Cendes, Y., Berger, E., Alexander, K. D., et al. 2022, , 938, 28

    2022

  9. [9]

    D., et al

    Cendes, Y., Berger, E., Alexander, K. D., et al. 2024, , 971, 185

    2024

  10. [10]

    A., 2010, , 187, 119

    Clarke D. A., 2010, , 187, 119

    2010

  11. [11]

    2019, , 483, 565

    Curd, B., & Narayan, R. 2019, , 483, 565

    2019

  12. [12]

    C., Roth, N., et al

    Dai, L., McKinney, J. C., Roth, N., et al. 2018, , 859, L20

    2018

  13. [13]

    2019, , 871, 126

    Gillessen, S., Plewa, P.M.., Widmann, F., et al. 2019, , 871, 126

    2019

  14. [14]

    J., van Velzen, S., Miller-Jones, J

    Goodwin, A. J., van Velzen, S., Miller-Jones, J. C. A., et al. 2022, , 511, 5328

    2022

  15. [15]

    J., Mummery, A., Laskar, T., et al

    Goodwin, A. J., Mummery, A., Laskar, T., et al. 2024, arXiv:2410.18665

    Pith/arXiv arXiv 2024

  16. [16]

    D., & Lifshitz, E

    Landau, L. D., & Lifshitz, E. M. 1987, Fluid Mechanics, 2nd edn. (Oxford: Elsevier Butterworth-Heinemann)

    1987

  17. [17]

    S., 1994, High Energy Astrophysics

    Longair M. S., 1994, High Energy Astrophysics. Volume 2. Stars, the Galaxy and the interstellar medium. Cambridge Univ. Press, Cambridge

    1994

  18. [18]

    2020, , 492, 686

    Lu, W., & Bonnerot, C. 2020, , 492, 686

    2020

  19. [19]

    Lu, W., Kumar, P., & Evans, N. J. 2016, , 458, 575

    2016

  20. [20]

    S., & Kallman, T

    Kara, E., Dai, L., Reynolds, C. S., & Kallman, T. 2018, , 474, 3593

    2018

  21. [21]

    2024, , 971, 49

    Matsumoto, T., & Piran, T. 2024, , 971, 49

    2024

  22. [22]

    & Wang, W

    Mou, G. & Wang, W. 2021, , 507, 1684

    2021

  23. [23]

    2021, , 908, 197

    Mou, G., Dou, L., Jiang, N., et al. 2021, , 908, 197

    2021

  24. [24]

    2022, , 510, 3650

    Mou, G., Wang, T., Wang W., Yang J. 2022, , 510, 3650

    2022

  25. [25]

    2025, arXiv: 2510.14715

    Mou, G. 2025, arXiv: 2510.14715

    Pith/arXiv arXiv 2025

  26. [26]

    2016, , 458, 4250

    Sadowski, A., Tejeda, E., Gafton, E., et al. 2016, , 458, 4250

    2016

  27. [27]

    C., arXiv: arXiv:2206.10641

    Steinberg, E., & Stone, N. C., arXiv: arXiv:2206.10641

    Pith/arXiv arXiv

  28. [28]

    L., Kwan T

    Thomsen L. L., Kwan T. M., Dai L., Wu S. C., Roth N., Ramirez-Ruiz E., 2022, , 937, L28

    2022

  29. [29]

    2016, PASJ, 68, L7

    Tsuboi M., Kitamura Y., Miyoshi M., et al. 2016, PASJ, 68, L7

    2016

  30. [30]

    J., Berezhko, E

    V \"o lk, H. J., Berezhko, E. G., & Ksenofontov, L. T. 2005, A&A, 433, 229

    2005

  31. [31]

    D., Li J., Russell C

    Wang Q. D., Li J., Russell C. M. P., Cuadra J., 2020, , 492, 2481

    2020

  32. [32]

    M., Zoghbi, A., et al

    Xiang, X., Miller, J. M., Zoghbi, A., et al. 2024, , 972, 106

    2024

  33. [33]

    D., Narayan, R., Quataert, E., et al

    Xu, Y. D., Narayan, R., Quataert, E., et al. 2006, , 640, 319

    2006

  34. [34]

    W., Mou, G., et al

    Yang, L., Shu, X. W., Mou, G., et al. arXiv:2509.21299

    arXiv

  35. [35]

    2024, , 962, L18

    Zhang, F., Shu, X., Yang, L., et al. 2024, , 962, L18

    2024

  36. [36]

    Zhuang, J., Shen, R.-F., Mou, G., Lu, W., 2025, , 979, 109

    2025