A Statistical Model of Magnetic Islands in a Large Current Layer

We develop a statistical model describing the dynamics of magnetic islands in very large current layers that develop in space plasma. Two parameters characterize the island distribution: the flux contained in the island and the area it encloses. We derive an integro-differential evolution equation for this distribution function, based on rules that govern the small-scale generation of secondary islands, the rates of island growth, and island merging. Our numerical solutions of this equation produce island distributions relevant to the magnetosphere and corona. We also derive and analytically solve a differential equation for large islands that explicitly shows the role merging plays in island growth.

💡 Research Summary

The paper presents a statistical framework for describing the evolution of magnetic islands (plasmoids) in very large current sheets that occur in space plasmas such as the Earth’s magnetosphere and the solar corona. Recognizing that direct kinetic (PIC) simulations cannot span the enormous scales involved (L ≈ 4000 d_i for the magnetopause, L ≈ 10⁶ d_i for coronal mass ejections) and that fluid (MHD) models miss the essential small‑scale physics of reconnection, the authors construct a kinetic‑like description in the reduced phase space defined by two island parameters: the magnetic flux ψ contained in the island and the geometric area A it encloses. The distribution function f(ψ, A, t) gives the number of islands per unit ψ and A, and its evolution captures the combined effects of island birth, growth by reconnection, convective loss, and binary merging.

Key ingredients of the model are:

1. Source term S(ψ, A) – a Gaussian injection representing the spontaneous formation of secondary islands at the electron‑ion skin‑depth scale near the X‑line.
2. Reconnection‑driven growth – a constant normalized reconnection rate ε≈0.1 leads to flux increase (\dot ψ = ε c_A B_0) and area increase (\dot A = 2 ε c_A \sqrt{π A}). This yields a characteristic island size r₀ = ε L when growth balances the convective transit time L/c_A.
3. Convective loss – islands are advected out of the system at the Alfvén speed, giving a sink term –c_A A/L · f.
4. Merging rules – when two islands collide, the resulting island’s area is the sum of the two areas, while its flux is the larger of the two parent fluxes (the smaller flux is reconnected away). The collision probability is proportional to a velocity v(ψ₁, A₁, ψ₂, A₂)Δt/L. The velocity interpolates between an Alfvénic hybrid speed for islands larger than the ion skin depth and a whistler‑driven scaling for sub‑electron‑scale islands. The explicit form (Eq. 1) ensures v→0 as r→0.

By analogy with the Boltzmann equation, the authors write a full integro‑differential evolution equation (Eq. 4) that includes the above four contributions. The left‑hand side contains the convective derivatives in ψ and A due to reconnection; the right‑hand side contains source, sink, and the gain/loss terms from binary mergers. Importantly, the merging terms conserve total magnetic area (∫A · (merging) dψ dA = 0), guaranteeing that the model respects the physical constraint that reconnection does not create or destroy magnetic flux, only redistributes it.

The paper first examines the limit where merging is negligible. Transforming to radius r = √(A/π) and defining F(ψ, r) = 2πr f(ψ, πr²), the authors obtain a linear transport equation (Eq. 6). With a delta‑function source, the steady‑state Green’s function solution (Eq. 7) shows that islands lie along the line ψ = B₀ r and decay exponentially with distance from the injection point. The characteristic scale r₀ = ε L emerges naturally; for magnetopause parameters (L ≈ 30 R_E, ε ≈ 0.1) this predicts islands of order a few Earth radii, consistent with THEMIS observations.

When the merging terms are retained, the full equation must be solved numerically. After nondimensionalizing time by L/c_A, flux by ε B₀ L, radius by ε L, and velocity by ε c_A, only one free parameter remains: the normalized source amplitude Sₙ = S_N ε L/(c_A L). By varying Sₙ over four orders of magnitude (4, 40, 400, 4000) the authors compute steady‑state distributions F_∞(ψ, r). Without merging, the distribution remains narrowly confined to ψ = B₀ r. With increasing S*ₙ, the merging terms dominate, causing the distribution to spread toward larger radii while the flux remains roughly unchanged. This reflects the physical picture that merging adds area but not flux, thereby reducing the in‑plane magnetic field B = ψ/r for large islands.

Moments of the kinetic equation provide additional insight. Integrating Eq. 4 over ψ and A yields an ordinary differential equation for the total island number N(t): \