WorksPhysics and Astronomy
Local relaxation and scale-dependent alignment in compressible, magnetized turbulence
Abstract
Driven net- and no-net-flux MHD turbulence simulations up to reveal sign-mixed velocity–magnetic, velocity–vorticity, and magnetic–current aligned patches below the energy equipartition scale. The first two angles scale as and , while the third angle varies weakly with scale. We develop and test a constant-flux transport model for departures from relaxed states, which predicts both exponents. These findings affect eddy anisotropy, reconnection-mediated turbulence onset, large-scale dynamos, and the nature of magnetized turbulence.
Introduction
Magnetized turbulence pervades laboratory and planetary plasmas [1, 2, 3], the solar wind and magnetosheath [4, 5], star-forming clouds and the interstellar medium [6, 7, 8, 9], and intergalactic and intracluster media [10, 11]. Its nonlinear MHD interactions regulate cascade rates, mixing, heating, reconnection, and the turbulent transport that sustains large-scale dynamos. Alignment weakens nonlinear interactions between velocity, , and Alfvénic magnetic fluctuations, . Counter-propagating Elsasser fluctuations, , interact on , so suppresses the Alfvénic nonlinearity [12, 13, 14, 15, 16, 17, 18, 19]. In guide-field turbulence, also controls the perpendicular eddy thickness, , relative to its length, [20, 21]. A scale-independent energy flux then gives the dynamic-alignment prediction and a spectrum [13, 14, 17, 22], with implications for reconnection-mediated turbulence [23, 24, 25, 26].
However, alignment also has a broader interpretation, in that it is a signature of plasma relaxation [5, 27, 28, 29, 30]. Relaxed MHD states are configurations in which nonlinear forces cancel, or become gradients, unable to drive further turbulent or nonlinear transfer. Three alignments are relevant when considering the MHD model. Alfvénic alignment (Alfvénization), , weakens induction and shear-Alfvén interactions; Taylor alignment (Taylorization), , weakens the Lorentz force; and Beltrami alignment (Beltramization), , makes the inertial term gradient-like. Here is the current density and is the vorticity. Each alignment corresponds directly to an ideal helicity invariant: cross helicity, ; current helicity, ; or kinetic helicity, . Classically, these aligned states arise by finding the global minimum-energy state of the flow, constrained by the helicities [5, 27, 28, 29, 30]. However, both in simulation and in the Earth’s magnetosheath the alignments are localized and patchy across space [5, 29, 31].
Recent work by Banerjee et al. [30] describes a theory for the relaxation of scale-dependent helicity and invariant correlations, named the principle of vanishing nonlinear transfer (PVNLT). In PVNLT, a relaxed state has vanishing nonlinear transfer of the invariant correlations (see Supplemental Material [32], Sec. III, for more details). The limit in which two-point helicity correlation functions become stationary gives
where is the total pressure and is a problem-dependent scalar potential. Thus gives the Alfvénic states, , while the magnetically dominated or pressure-balanced limits give Taylor alignment, ; when the Lorentz term is force-free or gradient-balanced, the same relaxation condition depletes the non-gradient part of , giving Beltrami alignment. In this study we show that departures from these relaxed states are themselves transferred through the cascade, which we use to model the scale-dependent alignment structure functions.
We show numerically and analytically that below the energy-equipartition scale , where Alfvén and eddy times are equal and rapid relaxation can occur [28, 29], eddies remain close to local relaxed states given by Eqs. (1) and (2). Here for no-net-flux and for net-flux. However, the large scales are not perfectly relaxed and they pass departures from Alfvénic and Beltrami states to smaller scales with an approximately constant flux, which we verify directly with shell-transfer measurements. For the measured subsonic scaling [9], constant downscale transport of the Alfvénic departure predicts . Because kinetic helicity is signed and globally cancels, the leading sign-even Beltrami departure is , giving . We test these predictions with two extreme-resolution, low plasma beta, , compressible MHD simulations performed on SuperMUC-NG, using more than core-hours: (1) a no-net-flux fluctuation-dynamo simulation [9] and (2) a strong-guide-field simulation. Both have and nearly three decades in resolved, inertial scales, spanning a large-scale, kinetically dominated supersonic cascade with and a subsonic, magnetically dominated cascade with [9]. This plasma parameter range spans most phases of interstellar turbu-
lence, weakly compressible space plasmas, and many astrophysical laboratory experiments [3, 33, 34, 35].
Numerical Simulations
We solve isothermal ideal MHD in a periodic domain with a hybrid-precision FLASH implementation and a positivity-preserving MUSCL-Hancock HLL5R scheme [36, 37, 38, 39, 40]. A finite-correlation-time, nonhelical Ornstein–Uhlenbeck force drives both simulations at [41, 42, 43, 44]. Their stationary states have , , and Alfvénic Mach number (defined using the rms field for the no-net-flux run and for the guide-field run); the latter has [45]. The End Matter section “Numerical details and Reynolds estimate” gives the initial conditions and effective .
The simulations have zero signed system-scale fluctuating helicities,
within statistical uncertainty. This excludes a single global relaxed state but permits sign-mixed local relaxation, for example,
where is signed. Such alignment can deplete induction and Lorentz forces and make the inertial term gradient-like while canceling in volume averages.
Patchy Alignment and Local Relaxation
Figure 1 shows that the compressible turbulence (left, no-net-flux; right, net-flux) organizes into extended patches with (top) and , separated by thin orthogonal interfaces, where the corresponding nonlinearities are largest. The white circles mark , the scale below which the cascade becomes magnetically dominated. Supplemental Material Figure S1 zooms into the guide-field slice and overlays the in-plane and streamlines inside an aligned patch and across an interface. The PDFs in Figure 2(a)–(Figure 2c) show that the same local organization also includes . We interpret the measured angles as distances from local relaxed states associated with the nonlinearities in the fluid equations, not only as geometric factors in an eddy-turnover time [9, 13]. This interpretation is motivated by PVNLT [30], where relaxed MHD states have vanishing nonlinear transfer of ideal invariant correlations. The nonlinear terms need not each vanish separately: they may vanish or be exactly balanced by gradients. Alfvénic, Taylor, and Beltrami alignments are the measurable limiting forms of this relaxation, with the observed – alignment signaling depleted or pressure-balanced Lorentz forces in both simulations.
Figure 1. Local alignment in no-net-flux and net-flux turbulence. Composite full-domain slices for (a) the no-net-flux simulation and (b) the net-flux guide-field simulation. In each column, the top half shows and the bottom half shows from the same slice; in (b), is directed out of the plane. The color encodes the local angle from parallel through perpendicular to anti-parallel alignment, with the lower color scale clipped at and for contrast. Dark interfaces mark nearly orthogonal regions where the corresponding cross products, or , are largest. White circles have radius , the energy-equipartition scale where the local Alfvén and eddy times are equal, with in (a) and in (b).
Figure 2. Global and scale-dependent alignment. (a)–(c) PDFs of , , and , with peaks near parallel and anti-parallel Alfvénic, Taylor, and Beltrami states [5, 29, 30, 46]. (d)–(f) Alignment structure functions of projected increments versus the strong-cascade separation : in the no-net-flux run and in the net-flux guide-field comparison, with increments projected perpendicular to the local two-point mean magnetic field before the angle is formed. Solid curves are no-net-flux statistics; dashed curves are net-flux guide-field statistics. Shaded bands show realization-to-realization scatter; vertical grey lines mark , , and the driving scale as labeled. Below , in both simulations and . The variation is small, rad () below , which is the sense in which the magnetic–current alignment is nearly scale independent; the guide-field curve remains compatible with a weak trend. The guide line is the dynamic-alignment prediction [13, 14].
Figure S1. Guide-field Alfvénic alignment. The color shows on an slice perpendicular to the imposed guide field, , in the net-flux simulation. The white boxes mark a strongly aligned patch and a nearby interface between aligned patches. The insets show the same regions with in-plane magnetic-field streamlines, , in white and in-plane velocity streamlines, , in gold.
Scale-Dependent Alignment
We measure first-order alignment structure functions at separation (definitions in the End Matter section “Definitions of scale-dependent alignment structure functions”). Figure 2(d)–(Figure 2f) shows the isotropic no-net-flux and perpendicular guide-field statistics defined in its caption. The sonic scale , where , is (no-net-flux) and (net-flux) [9, 40]. At the energy-equipartition scale , the scale-dependent Alfvén and eddy times are equal; smaller scales are magnetically dominated [9].
The scale-dependent range in the and structure functions begins at , which also corresponds to the characteristic size of the aligned patches in Figure 1. Below this scale, the no-net-flux run (solid lines) gives
while remains strongly aligned with only weak scale variation. The net-flux guide-field statistics (dashed lines) show the same shallow trend, offset to smaller angles, a comparable trend, and the same variation compatible with . In both simulations, the Alfvénic exponent is shallower than the strong-guide-field dynamic-alignment prediction [13, 14, 15]. Since the same scaling appears in both simulations when measured on the appropriate cascade scale, it is unlikely to be a peculiarity of the no-net-flux fluctuation dynamo. The scaling is, to our knowledge, the first measurement of scale-dependent Beltramization in MHD turbulence.
Because these structure functions are insensitive to the sign of the local alignment, they should be interpreted as average distances from the nearest parallel or anti-parallel PVNLT relaxed state. We established the sign-mixed character of those states by the real-space patches (Figure 1), the angle PDFs (Figure 2(a)–(Figure 2c)), and the vanishing signed helicity averages in Eq. (3), while the structure functions show how the corresponding unsigned departures decrease with scale.
These three measurements are captured by small departures from relaxed states. Consider the velocity and magnetic increment amplitudes scale as , e.g., on scales below . Using increment vectors and , the Alfvénic departure is the positive quadratic distance from ,
where the last inequality assumes nearly equipartitioned, small-angle increments. This difference norm is natural because Alfvénic relaxation requires equality, , not only parallelism. A scale-independent flux for carrying this first moment downscale, , gives
We test this prediction using the shell-to-shell departure transfer and its cross-scale flux (definitions in the End Matter section “Departure transfer diagnostics”). Figure 3 shows that closely tracks the energy flux, becomes approximately constant near , and is dominated by interactions between neighboring shells. We also compute total-energy transfers as references. The fields are filtered and downsampled before analysis; normalization, shell definitions, and conservative transfer operators are specified in End Matter. The shared constant-flux interval directly ties departure transport to the energy cascade.
Figure 3. Shell transfers and departure fluxes. (a),(c) Shell-to-shell matrices , from source shell to target shell , for total energy, , Alfvénic departure, , and Beltrami departure, , in the no-net-flux and net-flux simulations. Matrix colors are normalized by , the characteristic shell-transfer amplitude in the corresponding constant-flux interval. (b),(d) Downscale fluxes normalized by for each realization: (red), (blue), and (green). Shaded bands show realization scatter; grey bands mark the approximate constant-flux intervals; dashed lines mark . No-net-flux transfers use isotropic shells, , , and ; guide-field transfers use perpendicular shells, , , and .
For –, the analogous signed density is kinetic helicity. The End Matter section “Cross and kinetic helicity transport in the barotropic and force-free limits,” Eq. (26), shows that the kinetic helicity is made non-conservative by a non-gradient Lorentz force, but the observed Taylorization strongly suppresses that source. The remaining issue is sign: in the present nonhelical system the signed first moment can cancel between Beltrami and anti-Beltrami patches. The lowest useful sign-even Beltrami constraint is therefore the second moment of the Beltrami departure. With vorticity increment ,
which is angular because Beltrami states require proportionality between and , whose amplitudes and dimensions differ. The amplitude-weighted transfer diagnostic used to measure this transport is defined in the End Matter section “Departure transfer diagnostics,” Eqs. (34) and (35). The approximately constant intervals in Figure 3 support a scale-independent rate for carrying this second moment downscale, , which gives
Thus the different powers of the angle are not introduced symmetrically. The Alfvénic case has a positive first-moment departure inherited from the energy–cross-helicity square, while the kinetic-helicity case is signed and globally canceled, making a departure correlation the leading sign-even object. Supplemental Material, Sec. IV, gives a more detailed exponent derivation.
Discussion & Conclusions
Driven compressible MHD turbulence self-organizes into sign-mixed relaxed patches below , analogous to cross-helical patches in the solar wind and magnetosheath [5, 28, 47]. The measured shell transfers support a constant, local cascade of departures from these states: with , and , yield the observed Alfvénic and Beltrami exponents. Strong Taylorization suppresses the Lorentz-force source in the kinetic-helicity evolution, enabling conservative Beltrami-departure transport. Scale-dependent alignment sets eddy anisotropy: [13, 20, 21], so the measured gives rather than . This weaker sheet anisotropy changes the expected onset of a tearing-mediated cascade [23, 24, 48, 49]. For alignment and fluctuation-amplitude exponents and , the standard tearing balance gives
Thus gives , whereas the measured , gives . This shifts the tearing transition to smaller scales and substantially increases the Rm required to separate it from the resistive scale (Supplemental Material, Sec. II). Sign-mixed helicities also enter large-scale dynamo closures through the turbulent EMF, , residual kinetic–current helicity in , and cross helicity in [50, 51, 52, 53]. Here combines residual kinetic and current helicities, whereas is proportional to cross helicity; their cancellations can therefore differ across averaging scales and local environments. Alfvénic alignment suppresses the fluctuating cross product but enhances the scalar cross helicity supplying , which can dominate shear-layer dynamos [53]. It also sets a maximum averaging scale over which the signed helicities cancel, limiting the surviving scale-dependent dynamo contributions. MMS magnetosheath observations already show analogous relaxation-angle PDFs and cascade transfers [5, 54, 55]; measuring all three scale-dependent alignments would extend recent solar-wind – measurements [19].
Acknowledgments
J. R. B. acknowledges Christoph Federrath, whose help and guidance made it possible to perform the calculations at these scales. We acknowledge productive discussions with Bindesh Tripathi, Elias Most, Ralf S. Klessen, Salvatore Cielo, Uddipan Banik, Riddhi Bandyopadhyay, Eliot Quataert, Sasha Philippov, Philipp K. S. Kempski, Drummond Fielding, Alexander Chernoglazov, Hui Li, Philip Mocz, Bart Ripperda, and Chris Thompson. J. R. B. acknowledges high-performance computing resources provided by the Leibniz Rechenzentrum and the Gauss Centre for Supercomputing grants pr73fi, pn76gi, and pn76ga, and GCS large-scale project 10391. J. R. B. and A. B. acknowledge support from NSF Award 2206756. J. R. B. further acknowledges funding from the Natural Sciences and Engineering Research Council of Canada (NSERC) (funding reference number 568580). This work was in part supported by NASA through NASA Hubble Fellowship grant HST-HF2-51597.001-A, awarded by the Space Telescope Science Institute and administered by AURA for NASA under contract NAS5-26555.
End Matter
Numerical details and Reynolds estimate
To avoid spending the highest-resolution run on transient states, we initialize the largest grids with steady-state realizations from lower-resolution simulations. For both simulations, we drive boxes to stationarity, interpolate the steady-state turbulence to the grids, then drive for two more outer-scale turnover times , and average the main statistics, structure functions and PDFs, over 15 stationary realizations in the final . The forcing sets the target rms velocity ; in the net-flux run we then choose the imposed uniform field so that , matching the rms Alfvénic Mach number of the fluctuation-dynamo simulation. The simulations are implicit large-eddy simulations, with effective viscosity and resistivity supplied by the discretization. Calibrations against explicit-viscosity DNS give empirical conversions from grid scale to Re and Rm [56, 57, 58]. For supersonic MHD turbulence, Malvadi Shivakumar and Federrath [58] find with , , and with , . For , this gives and .
Definitions of scale-dependent alignment structure functions
To compute the scale-dependent alignment structure functions shown in Figure 2(d), we sample point pairs , where is position and is the lag vector. The scalar separation denotes the strong-cascade scale in each geometry. In the no-net-flux run,
The direction of is not fixed and is not conditioned on the local magnetic field. In the guide-field run, let be the imposed-field direction and define
In practice, following Mason et al. [14], the guide-field lags are sampled in the perpendicular plane, , so . For any vector fields and , we define raw increments in the bin as
with the appropriate lying in that bin. Following the standard local mean-field definition [18, 26], we define
and project the increments into the plane perpendicular to this two-point mean magnetic field,
We then construct the ratio of first-order structure functions
Here denotes an average over positions, sampled lag directions, and stationary-state realizations in the separation bin. The no-net-flux construction is applied to the pairs , , and ; the guide-field comparison applies the corresponding perpendicular-plane construction to the same three pairs. We restrict attention to this local-field projection because scale-dependent anisotropy and alignment of the strong MHD cascade are recovered in the local mean-field coordinate [14, 20, 25, 26, 59, 60]. Since the no-net-flux run has , it has no global parallel direction to exclude; the isotropic statistic is therefore the zero-guide-field analogue of the perpendicular guide-field statistic. Both choices identify the separation across which the strong cascade transfers fluctuation energy. We use sampled point pairs to ensure convergence of the structure functions on all scales [9, 40].
Cross and kinetic helicity transport in the barotropic and force-free limits
The ideal, barotropic, unforced (inertial) velocity, induction, and vorticity equations are
Here we define the Lorentz acceleration
which vanishes in the Taylor-relaxed limit, , and is gradient-like in the pressure-balanced limit, for scalar . We use ideal MHD with a barotropic plasma equation of state (including the isothermal equation solved here and general isentropic adiabatic plasmas), so the pressure force is a gradient,
which means the pressure term may enter through the conservative fluxes.
Helicity evolution equations
Taking gives
with and
The Lorentz acceleration drops out identically because , and the barotropic pressure force contributes as an enthalpy flux. Thus is a compressible, ideal invariant. Taking gives
where , with flux
In the absence of baroclinicity, , density enters only through the Lorentz acceleration. In the force-free case the kinetic-helicity source vanishes pointwise, while in the gradient-like case , so the source becomes a conservative flux and . Approximately conserved kinetic-helicity transport therefore requires a small non-gradient part of , and the measured – alignment is the stronger sufficient condition. Thus, in an ideal, unforced, force-free, barotropic but compressible plasma, both cross helicity, , and kinetic helicity, , are ideal invariants, justifying the structure functions studied in Figure 2.
Departure transfer diagnostics
We use Fourier shell filters , with and . No-net-flux transfers are binned in ; guide-field transfers are binned in and summed over . For any conservative shell transfer , the downscale flux through shell boundary is
We write for integrals over the periodic volume. Compressible advective transfers use
with the same definition for scalar . This operator is skew-adjoint for periodic boundaries, and the half-divergence term is required for velocity-mediated transfers in compressible flow [61]. The magnetic field below is the field used in the alignment statistic and cross helicity, up to a constant code-unit normalization; no local factor of is applied.
The total-energy transfer plotted in Figure 3 is
For Alfvénic departure, we define a fixed-sign departure and mediator fields
The fixed-sign shell transfer is
Figure 3 shows the sign-even matrix . For the flux, the nearest local Alfvénic branch is chosen at each cutoff,
The local sign mask makes a scale-local flux diagnostic rather than a conventional antisymmetric shell-transfer matrix.
For the Beltrami departure, we define the velocity component perpendicular to vorticity and the transported vector
with where is numerically zero, and , . Here is the velocity component that vanishes in either Beltrami branch, while the extra factor of makes , the amplitude-weighted, sign-even second-moment departure used above. Its conservative transfer and flux are
By construction, .
Supplemental Material for Local relaxation and scale-dependent alignment in compressible, magnetized turbulence
James R. Beattie; Amitava Bhattacharjee
Visualizations of local relaxed-state alignment
Figure S1 shows the guide-field simulation in real space, using the same angle as in the main text. Most of the slice is close to one of the two Alfvénic branches, or , while the dark, nearly orthogonal regions are comparatively thin interfaces where the Alfvénic nonlinearity is largest. In this sense, the visualization suggests that Alfvénic alignment is the volume-filling organization, with misalignment concentrated at boundaries between sign-mixed relaxed patches. The two zooms show an aligned region and a nearby interface; in both cases, the in-plane velocity and magnetic streamlines remain locally organized, consistent with the local-relaxation picture.
Figure S2 shows the corresponding full-slice organization of the three alignment angles used in the Letter. The Alfvénic, Beltrami, and Taylor angles all show extended aligned or anti-aligned regions separated by sharper, nearly orthogonal interfaces, reinforcing the interpretation that the measured structure functions quantify departures from locally relaxed states.
Figure S2. Three relaxed-state alignment angles. Composite no-net-flux slice showing the Alfvénic angle in the top partition, the Beltrami angle in the middle partition, and the Taylor angle in the bottom partition. The color encodes the local angle from parallel through perpendicular to anti-parallel alignment; the color scale is clipped at and for contrast.
Critical length scale for tearing-mediated cascade
Starting from the disruption criterion used in reconnection-mediated cascade models [23, 24, 48, 49], the critical eddy satisfies, up to the logarithmic Lambert- factor and order-unity constants,
Here and are the reconnecting-sheet thickness and length, is the Lundquist number based on , is the magnetic-fluctuation amplitude at , and is the magnetic diffusivity. Let be the alignment exponent and the magnetic-amplitude exponent, then,
The aligned eddy length and corresponding Lundquist number are then
Substituting into Eq. (36) gives
or
where is the critical wavenumber. For the usual dynamic-alignment scaling , this recovers
For the present alignment law, , together with the measured subsonic velocity scaling , Eq. (40) gives
Thus the weaker Alfvénic alignment shifts the tearing-mediated transition to smaller scales than the standard dynamic-alignment estimate. The same scaling also changes the Reynolds number needed for a resolved tearing-mediated interval. If the onset criterion is a fixed separation between the disruption scale and the resistive scale,
then
For the exponent in Eq. (44) changes from when to when . Therefore a critical value inferred in the standard dynamic-alignment case maps to
Taking , consistent with the threshold discussed by Dong et al. [26], gives , up to order-unity and logarithmic corrections.
PVNLT relaxed states
For an inviscid invariant built from fields and ,
PVNLT [30] considers the two-point correlator at separation ,
where the average is over position . The schematic evolution of the correlator is
Here is the nonlinear transfer of the correlator, while and denote forcing and dissipation contributions.
In a stationary inertial range, and , so the nonlinear transfer term must vanish,
This is the condition for relaxation in PVNLT. Applying Eq. (49) to incompressible MHD gives
for the energy, magnetic-helicity, and cross-helicity correlators, where is the total pressure built from plasma pressure and kinetic pressure , and are scalar fields [30]. Following Banerjee et al. [30], the aligned limits follow by taking the scalar-gradient part of the induction condition to be negligible. If , then
where the last relation follows for constant . Substitution into Eq. (50) gives
Thus the relaxed state is generally pressure- or gradient-balanced. In the low-, magnetically dominated limit discussed by Banerjee et al. [30], Eq. (53) reduces to , the Taylor or force-free limit. If instead , the same branch gives and Eq. (50) requires only a constant total pressure, the Alfvénic limit. Finally, Eq. (50) shows the Beltrami limit: if the Lorentz term is force-free or pressure-balanced, then the non-gradient part of is depleted, and the negligible-gradient limit gives . Thus PVNLT [30] naturally yields pressure- or gradient-balanced relaxed states, with Taylor, Alfvénic, and Beltrami alignments as measurable limiting cases.
ESTIMATE FOR THE ALIGNMENT EXPONENTS
The departure model keeps the PVNLT [30] idea that relaxation depletes nonlinear invariant transfer, but applies it at finite distance from a relaxed state. Let with in the measured subsonic range. For an Alfvénic state, the normalized positive distance from the nearest state is
for nearly equipartitioned, small-angle increments. If this first-moment departure is carried downscale at a scale-independent rate,
then
which gives the measured Alfvénic exponent. For the Beltrami departure, the local invariant density is kinetic helicity, . It is signed and globally canceled in the present nonhelical system, so the leading sign-even constraint is a second-moment departure. With
and a scale-independent rate for carrying this second moment downscale,
one obtains
The different powers follow from the invariant being transported: the Alfvénic alignment type has a positive first-moment distance built from energy and cross helicity, while the Beltrami alignment type is signed and therefore enters through a sign-even second moment. The End Matter section “Departure transfer diagnostics” gives the corresponding shell-transfer diagnostics.
References
- [1]C. A. Jones, Planetary magnetic fields and fluid dynamos, Annual Review of Fluid Mechanics 43, 583 (2011).DOI
- [2]A. S. Brun and M. K. Browning, Magnetism, dynamo action and the solar-stellar connection, Living Reviews in Solar Physics 14, 4 (2017).DOI
- [3]P. Tzeferacos, A. Rigby, A. F. A. Bott, A. R. Bell, R. Bingham, A. Casner, F. Cattaneo, E. M. Churazov, J. Emig, F. Fiuza, C. B. Forest, J. Foster, C. Graziani, J. Katz, M. Koenig, C. K. Li, J. Meinecke, R. Petrasso, H. S. Park, B. A. Remington, J. S. Ross, D. Ryu, D. Ryutov, T. G. White, B. Reville, F. Miniati, A. A. Schekochihin, D. Q. Lamb, D. H. Froula, and G. Gregori, Laboratory evidence of dynamo amplification of magnetic fields in a turbulent plasma, Nature Communications 9, 591 (2018).DOI
- [4]R. Bruno and V. Carbone, The solar wind as a turbulence laboratory, Living Reviews in Solar Physics 10, 2 (2013).
- [5]F. Pecora, Y. Yang, A. Chasapis, S. Servidio, M. E. Cuesta, S. Roy, R. Chhiber, R. Bandydapadhyay, D. J. Gershman, B. L. Giles, J. L. Burch, and W. H. Matthaeus, Relaxation of the turbulent magnetosheath, The Monthly Notices of The Royal Astronomical Society 525, 67 (2023), arXiv:2302.00634 [physics.flu-dyn].DOI
- [6]B. G. Elmegreen and J. Scalo, Interstellar Turbulence I: Observations and Processes, Annu. Rev. Astron. Astrophys. 42, 211 (2004), arXiv:0404451 [astro-ph].
- [7]R. S. Klessen, Star Formation in Molecular Clouds, in EAS Publications Series, EAS Publications Series, Vol. 51, edited by C. Charbonnel and T. Montmerle (2011) pp. 133–167, arXiv:1109.0467 [astro-ph.GA].DOI
- [8]B. M. Gaensler, M. Haverkorn, B. Burkhart, K. J. Newton-McGee, R. D. Ekers, A. Lazarian, N. M. McClure-Griffiths, T. Robishaw, J. M. Dickey, and A. J. Green, Low-Mach-number turbulence in interstellar gas revealed by radio polarization gradients, Nature (London) 478, 214 (2011), arXiv:1110.2896 [astro-ph.GA].DOI
- [9]J. R. Beattie, C. Federrath, R. S. Klessen, S. Cielo, and A. Bhattacharjee, The spectrum of magnetized turbulence in the interstellar medium, Nature Astronomy 9, 1195 (2025), arXiv:2504.07136 [astro-ph.GA].DOI
- [10]W. Schmidt, J. P. Schmidt, and P. Grete, Turbulence in the intragroup and circumgalactic medium, Astronomy and Astrophysics 654, A115 (2021), arXiv:2107.12125 [astro-ph.GA].DOI
- [11]R. Mohapatra and P. Sharma, Turbulence in the intracluster medium: simulations, observables, and thermodynamics, The Monthly Notices of The Royal Astronomical Society 484, 4881 (2019), arXiv:1810.00018 [astro-ph.GA].DOI
- [12]M. Dobrowolny, A. Mangeney, and P. Veltri, Properties of magnetohydrodynamic turbulence in the solar wind, Astronomy and Astrophysics 83, 26 (1980).
- [13]S. Boldyrev, Spectrum of magnetohydrodynamic turbulence, Physical Review Letters 96, 115002 (2006).pubmed.ncbi.nlm.nih.gov/16605832
- [14]J. Mason, F. Cattaneo, and S. Boldyrev, Dynamic Alignment in Driven Magnetohydrodynamic Turbulence, Phys. Rev. Lett. 97, 255002 (2006), arXiv:astroph/0602382 [astro-ph].pubmed.ncbi.nlm.nih.gov/17280362
- [15]J. C. Perez and S. Boldyrev, Role of Cross-Helicity in Magnetohydrodynamic Turbulence, Physical Review Letters 102, 025003 (2009), arXiv:0807.2635 [astro-ph].DOI
- [16]A. Mallet, A. A. Schekochihin, and B. D. G. Chandran, Refined critical balance in strong Alfvénic turbulence, The Monthly Notices of The Royal Astronomical Society 449, L77 (2015), arXiv:1406.5658 [astro-ph.SR].DOI
- [17]B. D. G. Chandran, A. A. Schekochihin, and A. A. Mallet, Intermittency and Alignment in Strong RMHD Turbulence, Astrophys. J. 807, 39 (2015), arXiv:1403.6354 [astro-ph.SR].arxiv.org/abs/1403.6354
- [18]A. Chernoglazov, B. Ripperda, and A. Philippov, Dynamic Alignment and Plasmoid Formation in Relativistic Magnetohydrodynamic Turbulence, The Astrophysical Journal Letters 923, L13 (2021), arXiv:2111.08188 [astro-ph.HE].DOI
- [19]N. Sioulas, M. Velli, A. Mallet, T. A. Bowen, B. D. G. Chandran, C. Shi, S. S. Cerri, I. J. Iodice, T. Ervin, and D. E. Larson, Scale-Dependent Dynamic Alignment in MHD Turbulence: Insights into Intermittency, Compressibility, and Imbalance Effects, arXiv e-prints , arXiv:2407.03649 (2024), arXiv:2407.03649 [physics.space-ph].arxiv.org/abs/2407.03649
- [20]A. Mallet, A. A. Schekochihin, B. D. G. Chandran, C. H. K. Chen, T. S. Horbury, R. T. Wicks, and C. C. Greenan, Measures of three-dimensional anisotropy and intermittency in strong Alfvénic turbulence, Monthly Notices of the Royal Astronomical Society 459, 2130 (2016).DOI
- [21]A. Mallet and A. A. Schekochihin, A statistical model of three-dimensional anisotropy and intermittency in strong Alfvénic turbulence, The Monthly Notices of The Royal Astronomical Society 466, 3918 (2017), arXiv:1606.00466 [physics.space-ph].DOI
- [22]J. C. Perez, J. Mason, S. Boldyrev, and F. Cattaneo, On the Energy Spectrum of Strong Magnetohydrodynamic Turbulence, Physical Review X 2, 041005 (2012), arXiv:1209.2011 [astro-ph.SR].arxiv.org/abs/1209.2011
- [23]N. F. Loureiro and S. Boldyrev, Role of Magnetic Reconnection in Magnetohydrodynamic Turbulence, Physical Review Letters 118, 245101 (2017), arXiv:1612.07266 [physics.plasm-ph].DOI
- [24]L. Comisso, Y. M. Huang, M. Lingam, E. Hirvijoki, and A. Bhattacharjee, Magnetohydrodynamic Turbulence in the Plasmoid-mediated Regime, The Astrophysical Journal 854, 103 (2018), arXiv:1802.02256 [physics.plasm-ph].DOI
- [25]C. Dong, L. Wang, Y.-M. Huang, L. Comisso, and A. Bhattacharjee, Role of the Plasmoid Instability in Magnetohydrodynamic Turbulence, Physical Review Letters 121, 165101 (2018), arXiv:1804.07361 [physics.plasm-ph].DOI
- [26]C. Dong, L. Wang, Y.-M. Huang, L. Comisso, T. A. Sandstrom, and A. Bhattacharjee, Reconnection-driven energy cascade in magnetohydrodynamic turbulence, Science Advances 8, eabn7627 (2022), arXiv:2210.10736 [astro-ph.SR].
- [27]J. B. Taylor, Relaxation of toroidal plasma and generation of reverse magnetic fields, Phys. Rev. Lett. 33, 1139 (1974).DOI
- [28]W. H. Matthaeus, A. Pouquet, P. D. Mininni, P. Dmitruk, and B. Breech, Rapid Alignment of Velocity and Magnetic Field in Magnetohydrodynamic Turbulence, Physical Review Letters 100, 085003 (2008), arXiv:0708.0801 [astro-ph].DOI
- [29]S. Servidio, W. H. Matthaeus, and P. Dmitruk, Depression of Nonlinearity in Decaying Isotropic MHD Turbulence, Phys. Rev. Lett. 100, 095005 (2008).pubmed.ncbi.nlm.nih.gov/18352719
- [30]S. Banerjee, A. Halder, and N. Pan, Universal turbulent relaxation of fluids and plasmas by the principle of vanishing nonlinear transfers, Physical Review E 107, L043201 (2023), arXiv:2209.12735 [physics.plasm-ph].DOI
- [31]D. N. Hosking and A. A. Schekochihin, Reconnection-controlled decay of magnetohydrodynamic turbulence and the role of invariants, Phys. Rev. X 11, 041005 (2021).
- [32]See Supplemental Material at [URL will be inserted by publisher] for visualizations of local relaxed-state alignment, the tearing-transition estimate, the PVNLT relaxed-state derivation, and the alignment-exponent derivation.
- [33]K. Ferrière, Plasma turbulence in the interstellar medium, Plasma Physics and Controlled Fusion 62, 014014 (2020), arXiv:1912.08237 [astro-ph.GA].DOI
- [34]R. Bandyopadhyay, J. R. Beattie, and A. Bhattacharjee, Density Fluctuation - Mach Number Scaling in Compressible, High Plasma Beta Turbulence: In Situ Space Observations and High - Reynolds Number Simulations, Astrophys. J. Lett. 982, L45 (2025), arXiv:2502.08883 [astro-ph.SR].arxiv.org/abs/2502.08883
- [35]S. Merlini, J. R. Beattie, and V. Valenzuela-Villaseca, Numerical simulations of shock-driven, supersonic turbulence in colliding three-temperature laboratory plasmas, arXiv e-prints , arXiv:2605.21672 (2026), arXiv:2605.21672 [physics.plasm-ph].arxiv.org/abs/2605.21672
- [36]B. Fryxell, K. Olson, P. Ricker, F. X. Timmes, M. Zingale, D. Q. Lamb, P. MacNeice, R. Rosner, J. W. Truran, and H. Tufo, FLASH: An Adaptive Mesh Hydrodynamics Code for Modeling Astrophysical Thermonuclear Flashes, The Astrophysical Journal Supplement 131, 273 (2000).
- [37]A. Dubey, R. Fisher, C. Graziani, G. C. Jordan, IV, D. Q. Lamb, L. B. Reid, P. Rich, D. Sheeler, D. Townsley, and K. Weide, Challenges of Extreme Computing using the FLASH code, in Numerical Modeling of Space Plasma Flows, Astronomical Society of the Pacific Conference Series, Vol. 385, edited by N. V. Pogorelov, E. Audit, and G. P. Zank (2008) p. 145.
- [38]F. Bouchut, C. Klingenberg, and K. Waagan, A multi-wave approximate riemann solver for ideal mhd based on relaxation ii: numerical implementation with 3 and 5 waves, Numerische Mathematik 115, 647 (2010).
- [39]K. Waagan, C. Federrath, and C. Klingenberg, A robust numerical scheme for highly compressible magnetohydrodynamics: Nonlinear stability, implementation and tests, Journal of Computational Physics 230, 3331 (2011), arXiv:1101.3007 [astro-ph.IM].DOI
- [40]C. Federrath, R. S. Klessen, L. Iapichino, and J. R. Beattie, The sonic scale of interstellar turbulence, Nature Astronomy 10.1038/s41550-020-01282-z (2021).DOI
- [41]V. Eswaran and S. B. Pope, An examination of forcing in direct numerical simulations of turbulence, Computers and Fluids 16, 257 (1988).DOI
- [42]W. Schmidt, C. Federrath, M. Hupp, S. Kern, and J. C. Niemeyer, Numerical simulations of compressively driven interstellar turbulence. I. Isothermal gas, Astronomy and Astrophysics 494, 127 (2009), arXiv:0809.1321 [astro-ph].DOI
- [43]C. Federrath, J. Roman-Duval, R. Klessen, W. Schmidt, and M. M. Mac Low, Comparing the statistics of interstellar turbulence in simulations and observations: Solenoidal versus compressive turbulence forcing, Astronomy and Astrophysics 512, 10.1051/0004-6361/200912437 (2010), arXiv:0905.1060.DOI
- [44]C. Federrath, J. Roman-Duval, R. S. Klessen, W. Schmidt, and M. M. Mac Low, TG: Turbulence Generator, Astrophysics Source Code Library, record ascl:2204.001 (2022), ascl:2204.001.
- [45]J. R. Beattie, C. Federrath, and A. Seta, Magnetic field fluctuations in anisotropic, supersonic turbulence, The Monthly Notices of The Royal Astronomical Society 498, 1593 (2020), arXiv:2007.13937 [astro-ph.GA].DOI
- [46]T. Stribling and W. H. Matthaeus, Relaxation processes in a low-order three-dimensional magnetohydrodynamics model, Physics of Fluids B: Plasma Physics 3, 1848 (1991), https://pubs.aip.org/aip/pfb/article-pdf/3/8/1848/1236776/1848_1_online.pdf.DOI
- [47]K. T. Osman, M. Wan, W. H. Matthaeus, B. Breech, and S. Oughton, Directional Alignment and Non-Gaussian Statistics in Solar Wind Turbulence, The Astrophysical Journal 741, 75 (2011).DOI
- [48]A. Mallet, A. A. Schekochihin, and B. D. G. Chandran, Disruption of sheet-like structures in Alfvénic turbulence by magnetic reconnection, The Monthly Notices of The Royal Astronomical Society 468, 4862 (2017), arXiv:1612.07604 [astro-ph.SR].
- [49]S. Boldyrev and N. F. Loureiro, Magnetohydrodynamic Turbulence Mediated by Reconnection, The Astrophysical Journal 844, 125 (2017), arXiv:1706.07139 [physics.plasm-ph].DOI
- [50]A. Pouquet, U. Frisch, and J. Léorat, Strong MHD helical turbulence and the nonlinear dynamo effect, Journal of Fluid Mechanics 77, 321 (1976).DOI
- [51]N. Yokoi, Unappreciated cross-helicity effects in plasma physics: anti-diffusion effects in dynamo and momentum transport, Reviews of Modern Plasma Physics 7, 33 (2023).arxiv.org/abs/2303.01834
- [52]A. Brandenburg, D. Elstner, Y. Masada, and V. Pipin, Turbulent processes and mean-field dynamo, Space Science Reviews 219, 55 (2023).DOI
- [53]B. Tripathi, A. E. Fraser, P. W. Terry, E. G. Zweibel, M. J. Pueschel, and R. Fan, Large-scale dynamos driven by shear-flow-induced jets, Nature 649, 848 (2026).DOI
- [54]R. Bandyopadhyay, A. Chasapis, R. Chhiber, T. N. Parashar, W. H. Matthaeus, M. A. Shay, B. A. Maruca, J. L. Burch, T. E. Moore, C. J. Pollock, B. L. Giles, W. R. Paterson, J. Dorelli, D. J. Gershman, R. B. Torbert, C. T. Russell, and R. J. Strangeway, Incompressive energy transfer in the earth’s magnetosheath: Magnetospheric multiscale observations, The Astrophysical Journal 866, 106 (2018), arXiv:1806.04275 [physics.space-ph].DOI
- [55]R. Bandyopadhyay, L. Sorriso-Valvo, A. Chasapis, P. Hellinger, W. H. Matthaeus, A. Verdini, S. Landi, L. Franci, L. Matteini, B. L. Giles, D. J. Gershman, T. E. Moore, C. J. Pollock, C. T. Russell, B. J. Strangeway, R. B. Torbert, and J. L. Burch, In situ observation of hall magnetohydrodynamic cascade in space plasma, Physical Review Letters 124, 225101 (2020), arXiv:1907.06802 [physics.space-ph].DOI
- [56]N. Kriel, J. R. Beattie, A. Seta, and C. Federrath, Fundamental scales in the kinematic phase of the turbulent dynamo, The Monthly Notices of The Royal Astronomical Society 513, 2457 (2022), arXiv:2204.00828 [astro-ph.SR].arxiv.org/abs/2204.00828
- [57]P. Grete, B. W. O’Shea, and K. Beckwith, As a Matter of Dynamical Range - Scale Dependent Energy Dynamics in MHD Turbulence, The Astrophysical Journal Letters 942, L34 (2023), arXiv:2211.09750 [astro-ph.GA].DOI
- [58]L. Malvadi Shivakumar and C. Federrath, Numerical viscosity and resistivity in MHD turbulence simulations, arXiv: e-prints arXiv:2311.10350 (2023), arXiv:2311.10350 [astro-ph.SR].DOI
- [59]J. Cho and E. T. Vishniac, The anisotropy of magnetohydrodynamic alfvenic turbulence, The Astrophysical Journal 593, 273 (2000).DOI
- [60]J. Maron and P. Goldreich, Simulations of Incompressible Magnetohydrodynamic Turbulence, The Astrophysical Journal 554, 1175 (2001), astro-ph/0012491.DOI
- [61]P. Grete, B. W. O’Shea, K. Beckwith, W. Schmidt, and A. Christlieb, Energy transfer in compressible magnetohydrodynamic turbulence, Physics of Plasmas 24, 092311 (2017), arXiv:1706.06339 [physics.plasm-ph].DOI