Propagating Fluctuation Model Overview

• Propagating fluctuation models are frameworks that define how fluctuations are generated, transported, and transformed via explicit dynamical mechanisms across diverse systems.
• Analyses reveal that varying propagation mechanisms—ballistic, diffusive, or mode-coupled—across settings like disordered magnets, biochemical networks, and plasma turbulence yield distinct diagnostic signatures.
• Diagnostic tools such as structure factors, power spectral density breaks, and fluctuation decompositions are central to validating these models and connecting theory with observable phenomena.

Taken together, the literature suggests that “propagating fluctuation model” functions as an umbrella concept rather than a single canonical formalism. Across disordered magnets, stochastic dynamical systems, biochemical networks, plasma turbulence, heavy-ion collisions, and accretion theory, the common idea is that fluctuations are not merely local noise sources: they are transported, amplified, damped, or converted by an explicit dynamical mechanism, and the resulting propagation law becomes part of the model itself [(Acharyya, 2012); (Roy et al., 2023); (Ahmad et al., 2014); (Hogg et al., 2015); (Turner et al., 2021); (Nies et al., 2024)].

1. Scope and recurring structure

A recurring pattern in these works is the coexistence of four elements: a carrier of fluctuations, a propagation mechanism, a medium that filters or transforms the fluctuations, and an observable used to diagnose what survives. The carrier may be a traveling magnetic field, a front position, an empirical-measure fluctuation field, a zonal mode, or an inwardly propagating accretion-rate perturbation. The propagation mechanism may be ballistic, diffusive, Markovian, mean-field, or mode-coupled. The medium may be quenched disorder, a nonlinear promoter logic, a turbulent plasma, a hydrodynamic stage, or a viscous disc. The observable may be a structure factor, a coefficient of variation, a large-deviation function, a heat flux, a PSD, or a front diffusion coefficient.

Setting	Propagating quantity	Diagnostic
RFIM with traveling drive (Acharyya, 2012)	$$h_p(y,t)=h_0\cos\!\left(2\pi f t-\frac{2\pi y}{\lambda}\right)$$	$|S(k,t)|$
Coherent feed-forward loop (Roy et al., 2023)	one-step, two-step, and cross fluctuation terms	$\eta_y^2$
Stochastic point vortices (Shao et al., 12 Jan 2025)	$\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$	linear SPDE limit
Stochastic accretion disc (Turner et al., 2021)	stochastic $\alpha$ and inward $\dot M$ fluctuations	PSD break, rms–flux
Tokamak turbulence (Nies et al., 2024)	toroidal secondary zonal flow	$\langle Q\rangle/Q_{\mathrm gB}$, $E^{\mathrm ZF}$

This suggests that the term is best understood operationally: a propagating fluctuation model specifies not only where fluctuations are generated but also how they move through the system and how their transport changes the macroscopic state.

2. Spatially propagating drives, transport, and fronts

One concrete realization is the two-dimensional random-field Ising ferromagnet driven by a propagating magnetic field wave. Its Hamiltonian includes a quenched bimodal random field and an external drive

$$h_p(y,t)=h_0\cos\!\left(2\pi f t-\frac{2\pi y}{\lambda}\right),$$

so the field itself is a traveling modulation along the $y$-direction. Under zero-temperature single-spin-flip Monte Carlo dynamics, four nonequilibrium steady states were identified: a pinned all-up state, a propagating coherent strip phase with sharp boundaries, a propagating strip phase with rough boundaries, and a disordered non-propagating phase. These are distinguished by the line-magnetization structure factor $|S(k,t)|$0, with phase 2 characterized by $|S(k,t)|$1, phase 3 by $|S(k,t)|$2, and phases 1 and 4 by vanishing or near-vanishing $|S(k,t)|$3. The phase diagram in the $|S(k,t)|$4 plane is empirical, and the work explicitly notes that it does not determine whether the boundaries are sharp phase transitions or crossovers (Acharyya, 2012).

A different spatial transport problem appears in the tight-binding wave packet propagating in a fluctuating periodic potential. There the microscopic dynamics is

$|S(k,t)|$5

with the temporal fluctuations supplied by a stationary Markov process. Under the paper’s sectoriality and spectral-gap assumptions for the Markov generator, the square amplitude after diffusive rescaling converges not to a single heat kernel but to a superposition of heat-equation solutions indexed by quasi-momentum. The limiting macroscopic density is therefore diffusive, but only after the Bloch-periodic structure is retained through a quasi-momentum integral (Hamza et al., 2010).

Classical particle transfer between two reservoirs gives a further variant. The joint process $|S(k,t)|$6 obeys a tilted evolution operator whose spectrum is discrete once the particle-number distribution decays faster than exponentially. A special propagating initial condition satisfying $|S(k,t)|$7 for all $|S(k,t)|$8 excites only the lowest eigenmode of that operator, so the current statistics approach the large-deviation regime in a particularly transparent way. In that setting the propagating object is not a front or wave but the current-generating function itself under time evolution, and the steady-state fluctuation theorem follows from analyticity and the discrete spectrum of the tilted generator (Harbola et al., 2013).

3. Cascades, front wandering, and fluctuation fields

A more abstract propagating fluctuation model is the stochastic cascade amplification scenario. The basic stochastic differential equation

$|S(k,t)|$9

describes a stable attractor perturbed by additive noise, while the linearized separation obeys

$\eta_y^2$0

Although the average Lyapunov exponent is negative, transient intervals with $\eta_y^2$1 multiplicatively amplify small perturbations. In the tail regime this yields

$\eta_y^2$2

and the exponent $\eta_y^2$3 is interpreted as a negative fractal dimension through $\eta_y^2$4. The fluctuation is therefore propagated not by ordinary diffusion but by a stochastic sequence of amplification and decay episodes (Wilkinson et al., 2015).

Reaction–diffusion fronts provide a complementary picture. In the deterministic bistable equation

$\eta_y^2$5

the sign of the front velocity is controlled by the effective potential difference $\eta_y^2$6. In the spruce-budworm example, however, the stochastic birth–death–diffusion process introduces a different quasipotential $\eta_y^2$7, and the paper shows that fluctuations can reverse the direction of propagation by changing the relative stochastic stability of the two phases. For propagating population fronts into metastable states, a weak-noise theory yields diffusive front motion with

$\eta_y^2$8

while a WKB analysis determines the probability distribution of rare, large fluctuations of the front position and the most likely density profile associated with a prescribed average front velocity [(Khain et al., 2010); (Meerson et al., 2011)].

At the mean-field level, the stochastic point vortex model with common noise turns the propagation-of-fluctuations problem into a central-limit theorem for empirical measures. With

$\eta_y^2$9

the fluctuation field converges to the unique probabilistically strong solution of a linear stochastic evolution equation around the conditional McKean–Vlasov limit. The limiting equation contains both additive conditionally Gaussian noise and multiplicative transport noise $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$0. Because the covariance depends on the random mean-field density $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$1, the unconditional law is typically a mixture of Gaussians rather than Gaussian. Here propagation is linearized transport of fluctuations around a random mean field, not propagation of a single front or mode (Shao et al., 12 Jan 2025).

4. Pathway-based and mode-coupled propagation

In biochemical network theory, propagation is often decomposed by pathway. For coherent feed-forward loops with nodes $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$2, $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$3, and $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$4, the linear-noise analysis around steady state gives

$\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$5

or equivalently

$\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$6

The four terms represent intrinsic noise of $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$7, the direct one-step pathway $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$8, the indirect two-step pathway $\eta_t^N=\sqrt{N}\,(\mu_N-v_t)$9, and the cross-interaction between pathways. Additive and multiplicative integration at the $\alpha$0 promoter alter the derivatives $\alpha$1, $\alpha$2, and $\alpha$3, and therefore alter which coherent feed-forward loop motifs are degenerate or non-degenerate in each fluctuation channel. The total output noise is non-degenerate across C1–C4 even when individual components are degenerate, and C1-FFL displays the maximal level of fluctuations under the standard parameter set, especially for multiplicative integration (Roy et al., 2023).

A structurally similar, but physically different, reduced-order model appears in annular-mode dynamics. There the state is the pair of zonal indices $\alpha$4 associated with EOF1 and EOF2 of the zonal-mean zonal wind, and the eddy forcing is written with both self-feedback and cross-EOF feedback: $\alpha$5 The propagating regime is the decaying-oscillatory case in which the eigenvalues are complex, with existence condition

$\alpha$6

which in particular requires $\alpha$7. Stronger cross-EOF feedbacks increase the oscillation frequency and reduce the persistence of the annular modes. Application of the coupled-EOF model to Southern Hemisphere reanalysis data shows strong cross-EOF feedbacks, so the propagating annular mode is not reducible to a single red-noise EOF1 process (Lubis et al., 2020).

These pathway and mode-coupled formulations make explicit that propagation need not be spatial. It can instead mean transmission between parallel regulatory routes or between coupled large-scale modes, provided the fluctuation budget is decomposed and the cross terms are retained.

5. Plasma, collisionless shocks, and propagating zonal modes

Hybrid simulations of parallel propagating Alfvénic fluctuations give a kinetic realization in which the propagating structures are phase-steepened fronts. Starting from a large-amplitude, constant-$\alpha$8, broad-band Alfvénic fluctuation with $\alpha$9, $\dot M$0, and $\dot M$1, dispersive effects drive phase steepening and generate localized compressive fluctuations and parallel electric fields. Proton scattering at the steepened edges causes non-adiabatic perpendicular heating, while the parallel electric field at the fronts accelerates protons along the mean field. The asymptotic state contains a field-aligned beam at the Alfvén speed and greatly reduced compressibility, so the front itself is the fluctuation carrier that mediates energy transfer from wave to particles (González et al., 2021).

Relativistic Weibel-mediated shocks propagating into an inhomogeneous pair plasma show a related transfer from upstream fluctuations to downstream structures. With an imposed upstream density modulation

$\dot M$2

the shock converts the incoming inhomogeneity into downstream sound waves and entropy waves. These modes survive because the estimated diffusive decay times greatly exceed the simulation time, and the resulting temperature anisotropy provides free energy large enough to explain the observed field strength. Relative to the homogeneous case, the downstream magnetic field decays more slowly, so the fluctuating upstream medium imprints long-lived downstream magnetic structure (Tomita et al., 2019).

In magnetised tokamak turbulence, the propagating fluctuation is the toroidal secondary zonal flow. This zonal mode grows and propagates due to the combined effects of zonal-flow shearing and advection by the magnetic drift, with frequency $\dot M$3. Above a threshold in turbulence amplitude, small-scale toroidal secondary modes become unstable and shear apart turbulent eddies, forcing the turbulence level to remain near threshold. Incorporating this physics into a critical-balance saturation theory yields

$\dot M$4

with a corresponding scaling for the zonal-flow energy. Here propagation is inseparable from saturation: the zonal mode is both a transported fluctuation and the regulator of the turbulent state (Nies et al., 2024).

6. Propagation through evolving media: heavy-ion collisions and accretion flows

In relativistic heavy-ion phenomenology, propagation refers to the survival of event-by-event fluctuations through a sequence of nonequilibrium stages. Using the hybrid version of UrQMD, fluctuations of multiplicities, their ratios, and net-charge or net-proton observables were followed from early times through hadronic transport, optional hydrodynamic evolution, particlization, and the hadronic afterburner. Two standard measures,

$\dot M$5

are substantially reduced at freeze-out: dominant structures present at the initial stage of the evolution get smoothen out. Yet the energy dependence remains preserved till the freezeout, and the hydrodynamic evolution with chiral equation of state shows considerably higher fluctuation at lower collision energy than either pure transport or the hybrid with hadronic equation of state. The propagating fluctuation model here is therefore a dynamical baseline for what fluctuation signatures can survive through realistic medium evolution (Ahmad et al., 2014).

Accretion-disc variability supplies perhaps the most widely used contemporary meaning of the term. In the global three-dimensional MHD thin-disc simulation with $\dot M$6, dynamical-timescale turbulent Maxwell-stress fluctuations are too rapid to produce radially coherent accretion-rate variations, but low-frequency quasi-periodic dynamo action introduces low-frequency fluctuations in the Maxwell stresses which then drive the propagating fluctuations. Mass accretion rate and emission proxies recover log-normality, linear RMS-flux relations, and radial coherence consistent with inter-band lags, thereby linking phenomenological propagating-fluctuation models to MRI turbulence and dynamo-modulated stresses (Hogg et al., 2015).

The one-dimensional stochastic $\dot M$7-disc calculations sharpen this picture by solving the nonlinear diffusion equation with a stochastic viscosity field. These simulations find propagating fluctuations across a wide range of model parameters, explicitly contradicting previous work that had suggested otherwise. The most important control parameter is the timescale on which the viscosity fluctuations occur, and the paper reports a fitting formula for the thermal-disc PSD break: $\dot M$8 with $\dot M$9, $\langle Q\rangle/Q_{\mathrm gB}$0, and $\langle Q\rangle/Q_{\mathrm gB}$1. This directly connects an observable timing feature to the fluctuation timescale of the effective viscosity and, by implication, to the underlying MRI dynamo physics (Turner et al., 2021).

7. Diagnostics, limitations, and open questions

Across these literatures, diagnostics are unusually central to the definition of the model itself. In driven magnets the wavelength-resolved structure factor distinguishes coherent propagation from disorder (Acharyya, 2012). In biochemical networks the physically meaningful object is often not a single variance but its decomposition into intrinsic, direct, indirect, and cross-interaction components (Roy et al., 2023). In point-vortex theory the object of interest is a fluctuation field in $\langle Q\rangle/Q_{\mathrm gB}$2 rather than a scalar observable, and its limit is identified through martingale methods and a linear SPDE (Shao et al., 12 Jan 2025). In accretion theory the characteristic signatures are log-normal fluxes, linear rms–flux relations, coherence, lags, and PSD breaks (Hogg et al., 2015, Turner et al., 2021). In plasma turbulence the decisive evidence lies in spectral transfer, growth-rate thresholds, and transport scaling (Nies et al., 2024).

Several recurrent limitations are also explicit. The RFIM study identifies phase boundaries empirically and does not provide finite-size scaling or critical exponents (Acharyya, 2012). The coherent feed-forward loop analysis is based on linearization around steady state and notes that for stronger nonlinearities, such as $\langle Q\rangle/Q_{\mathrm gB}$3, the theory and simulations diverge (Roy et al., 2023). The UrQMD heavy-ion baseline contains no explicit critical dynamics and uses ideal rather than fluctuating hydrodynamics (Ahmad et al., 2014). The point-vortex fluctuation theory handles common noise rigorously, but the resulting unconditional fluctuation law is typically non-Gaussian because it is a mixture over the random environment (Shao et al., 12 Jan 2025). The Alfvénic-front and Weibel-shock simulations omit electron kinetic physics or full three-dimensionality, both of which are identified as important extensions (González et al., 2021, Tomita et al., 2019). The accretion-disc studies still require radiation MHD, broader radial domains, and fully relativistic treatments for quantitative comparison with spectral-timing data (Hogg et al., 2015, Turner et al., 2021).

A plausible implication is that future work on propagating fluctuation models will continue to move in two directions at once. One direction is rigorous closure: stronger mean-field limits, fluctuation SPDEs, and large-deviation theories. The other is richer physics: finite temperature in driven disordered systems, explicit critical dynamics in heavy-ion matter, kinetic-electron and three-dimensional effects in plasma turbulence, and radiation-coupled MHD in accretion flows. What remains stable across these domains is the central methodological claim: to understand fluctuations, one must model not only how they are generated, but how they propagate.