On the construction of general large-amplitude spherically polarised Alfvén waves

Problem Statement
The paper addresses the challenge of constructing general, large-amplitude, spherically polarised Alfvén wave solutions in a magnetised plasma. While the existence of these nonlinear solutions is well-established for small amplitudes (where $\delta B \ll B$) and in one dimension, there is no general method to construct them when the fluctuation amplitude is comparable to or exceeds the mean field ($\delta B \gtrsim |B|$), particularly in two or three dimensions. The core difficulty lies in satisfying two simultaneous constraints: the magnetic field must be divergence-free ($\nabla \cdot \mathbf{B} = 0$) and maintain a constant magnitude ($|\mathbf{B}| = \text{const.}$). These twin constraints leave only a single degree of freedom to define the 3-D vector field $\mathbf{B}$. Previous attempts to construct such solutions often resulted in spurious discontinuities or were limited to specific, restricted geometries because they attempted to specify one component of the field and solve for the others, a process that fails to guarantee real, smooth solutions as amplitudes grow.

Methodology
The authors propose a computational method that circumvents the difficulty of directly solving for a constrained high-amplitude field by "growing" a solution from a small-amplitude, linear Alfvénic fluctuation. The approach involves:

1. Splitting the Field: The magnetic field is decomposed into a mean component ($\mathbf{B}$) and a fluctuating component ($\delta \mathbf{B}$).
2. Evolution Equation: The fluctuation $\delta \mathbf{B}$ is evolved in time using a modified induction equation with an exponential growth term:
   $$\frac{\partial \delta \mathbf{B}}{\partial t} = \delta \mathbf{B} + \nabla \times [\tilde{\mathbf{u}} \times (\mathbf{B} + \delta \mathbf{B})]$$
   Here, $\tilde{\mathbf{u}}$ is a non-Alfvénic, compressive flow (specifically a potential flow $\tilde{\mathbf{u}} = \nabla \phi$) introduced solely to reshape the field as it grows, ensuring the total field magnitude remains constant.
3. Constraint Enforcement: To maintain $|\mathbf{B} + \delta \mathbf{B}|^2 = \text{const.}$, the potential $\phi$ is determined by solving a Poisson-like equation derived from the condition that the fluctuating part of the time derivative of $B^2$ is zero:
   $$B^2 \nabla^2_\perp \phi = -\delta \mathbf{B} \cdot \mathbf{B}$$
   where $\nabla_\perp$ represents the gradient operator perpendicular to the local field.
4. Initial Conditions: The process begins with a low-amplitude solution constructed using vector potentials (e.g., $\delta \mathbf{B} = \nabla \times (A_x \hat{x})$) and optimization techniques (such as MATLAB's fsolve or lsqnonlin) to ensure the initial state satisfies the constant magnitude constraint. The solution is then evolved numerically using finite-difference methods.

Key Contributions and Results
The paper presents the first general computational framework for generating these solutions in arbitrary dimensions and explores their properties:

• One-Dimensional (1-D) Solutions: The method successfully reproduces known 1-D solutions. As the amplitude grows ($A \equiv (\delta B^2/B^2)^{1/2} \gg 1$), the solutions approach a zero-mean-field limit. Crucially, 1-D solutions remain smooth and do not develop discontinuities, even at very high amplitudes ($A \approx 400$).
• Two-Dimensional (2-D) Solutions: The authors demonstrate that smooth, large-amplitude 2-D solutions exist (e.g., Fig. 3). However, the results show a strong dependence on the initial low-amplitude structure. In many cases, as the amplitude increases, the solutions develop extremely sharp gradients or discontinuities (e.g., Fig. 2). The paper notes that while some 2-D solutions converge at moderate resolutions, others show no sign of convergence even at high resolutions ($384^2$), suggesting the formation of true discontinuities.
• Three-Dimensional (3-D) Solutions: The method is extended to 3-D. Similar to the 2-D case, the solutions tend to develop sharp field structures. The authors note that while their current non-parallelised implementation limits resolution, the behavior is consistent with the 2-D findings: discontinuities form readily in multi-dimensional geometries unless the solution is effectively one-dimensional.
• Gradient Analysis: A scan of resolution versus solution amplitude (Fig. 5) reveals that 1-D solutions converge quickly, whereas 2-D solutions often exhibit non-convergent behavior indicative of discontinuity formation. The authors suggest that while smooth solutions likely exist for general initial conditions, they may require extremely sharp gradients to satisfy the constant-magnitude constraint in higher dimensions.

Significance and Claims
The authors claim their work provides a natural explanation for the extremely sharp field discontinuities observed in solar wind "switchbacks" (magnetic field reversals). The method mimics the physical process of wave growth in the expanding solar wind, where small-amplitude fluctuations grow as they propagate outward.

• Physical Relevance: The results suggest that while 1-D models (previously used to study switchbacks) predict smooth evolution, the more general 2-D and 3-D nature of solar wind turbulence naturally leads to the formation of sharp gradients and discontinuities as amplitudes grow. This offers a potential resolution to the difficulty 1-D models had in explaining the sharpness of observed switchbacks.
• Utility: The constructed solutions serve as valuable initial conditions for other numerical studies, particularly those investigating parametric decay and reflection-driven turbulence, which have previously been limited by a lack of general, large-amplitude Alfvénic inputs.
• Mathematical Insight: The paper provides a method to construct a general divergence-free, unit-vector field, complementing other classes of MHD solutions like force-free magnetostatic equilibria.

The authors remain modest regarding the mathematical completeness of the method, acknowledging that the Poisson-like equation used to determine the flow $\phi$ may lack exact solutions at finite resolution in complex geometries, and that the specific conditions leading to discontinuity formation versus smooth solutions require further study.