Scaling Limits for Beam Wave Propagation in Atmospheric Turbulence

Albert Fannjiang, Knut Sølna

Abstract

We prove the convergence of the solutions of the parabolic wave equation to that of the Gaussian white-noise model widely used in the physical literature. The random medium is isotropic and is assumed to have integrable correlation coefficient in the propagation direction. We discuss the limits of vanishing inner scale and divergent outer scale of the turbulent medium.

Introduction

When a laser or microwave beam propagates through the atmosphere, it encounters refractive turbulence caused by fluctuations in temperature, pressure, and humidity. These fluctuations manifest as three primary challenges for optical engineering: beam wander (displacement from the target), beam broadening (spatial spreading), and scintillation (intensity fluctuations).

In physical literature, these phenomena are frequently modeled using a Gaussian white-noise approximation. While this simplification makes the complex mathematics of wave propagation more tractable, it has historically lacked a rigorous mathematical foundation. The actual atmosphere is characterized by multi-scale structures described by Kolmogorov turbulence theory, not idealized white noise. This paper bridges that gap by providing a formal proof that the parabolic wave equation converges to the Gaussian white-noise model under specific scaling limits as the propagation distance increases.

Theoretical Framework and Methodology

The research utilizes a sophisticated toolkit from stochastic partial differential equations (SPDEs) to establish this convergence. The authors begin with the scalar Helmholtz equation, transitioning to the parabolic wave equation—a valid approximation when backscattering is negligible, as is typical for directed energy beams.

Non-dimensional Scaling

To analyze the behavior of the wave, the authors introduce three critical non-dimensional parameters:

  • $\varepsilon$: The ratio of the Fresnel length to the propagation distance.
  • $\eta$: The ratio of the Fresnel length to the "outer scale" $L_0$ (representing the largest turbulent eddies).
  • $\rho$: The ratio of the Fresnel length to the "inner scale" $\ell_0$ (representing the smallest eddies where dissipation occurs).

The Martingale Approach

The core of the proof relies on a martingale formulation. By defining an infinitesimal operator for the wave process, the authors demonstrate that as $\varepsilon \to 0$, the physical process (the "shaky" air) converges in distribution to a diffusion process. This diffusion process corresponds exactly to the Gaussian white-noise model used by physicists. This approach allows for testing against realistic turbulent models, such as the modified Von Kármán spectrum and the Hill spectrum, the latter of which accounts for the characteristic "bump" in turbulence intensity near the inner scale.

Key Insights and Mathematical Results

The White-Noise Limit

The study reveals that the white-noise limit is a natural consequence of "weak fluctuations" accumulated over long distances. As the propagation distance $L_z$ significantly exceeds the Fresnel length, the cumulative effect of numerous small-scale turbulent interactions converges into a single Gaussian process.

Handling Scale Extremes

The proof is robust enough to handle the complexities of the Hill spectrum, which is vital for real-world applications where intensity fluctuations show specific peaks near the inner scale. Furthermore, the authors address the mathematical regularity of the limit. They show that when the inner scale vanishes ($\rho \to \infty$), the resulting white-noise velocity field possesses a specific Hölder regularity ($H + 1/2$), providing a mathematical bound on the "roughness" of the model.

In scenarios where the outer scale becomes infinite ($\eta \to 0$), large-scale fluctuations can become mathematically uncontrollable. The authors provide a method to "factor out" these fluctuations, ensuring a stable and unique solution.

Core Results

  1. Theorem 1: The primary result proves that the solution to the rescaled parabolic wave equation converges to the Gaussian Markovian model.
  2. Uniqueness: The authors establish the uniqueness of the solution for the $n$-point correlation function, ensuring the model remains physically predictive.
  3. Inertial Range Definition: The paper rigorously defines beam behavior as the inner scale $\ell_0$ approaches zero and the outer scale $L_0$ approaches infinity, offering a complete mathematical picture of the inertial range of turbulence.

Discussion and Limitations

While this paper provides a rigorous foundation for the white-noise model, it operates under the assumption of an isotropic medium (uniform in all directions). In practical atmospheric conditions—particularly near the ground or in high-wind environments—turbulence can be anisotropic. Additionally, the proof assumes an integrable correlation coefficient in the direction of propagation.

Future research may need to incorporate more granular atmospheric data to account for refractive index fields outside the inertial range, especially for wave statistics that are highly sensitive to the exact values of the inner and outer scales.

Summary updated

Abstract

We prove the convergence of the solutions of the parabolic wave equation to that of the Gaussian white-noise model widely used in the physical literature. The random medium is isotropic and is assumed to have integrable correlation coefficient in the propagation direction. We discuss the limits of vanishing inner scale and divergent outer scale of the turbulent medium.
BibTeX
@inproceedings{fannjiang2004scaling,
  author = {Albert Fannjiang and Knut Sølna},
  title = {{Scaling Limits for Beam Wave Propagation in Atmospheric Turbulence}},
  booktitle = {Stochastics},
  year = {2004},
  doi = {10.1142/s0219493704000973},
}
Figure