Introduction to Light Scattering

Light scattering in the atmosphere governs many of the visual and climatic phenomena we observe daily—from the deep blue of a clear midday sky to the rich reds of sunrise and sunset. At a fundamental level, scattering occurs when a beam of electromagnetic radiation encounters a particle or molecule and is redirected into many directions. Understanding this process is essential not only for explaining everyday optical effects but also for interpreting remote sensing data, improving climate models, and designing optical instruments.

The nature of the scattering depends on several key parameters: the size of the scatterer relative to the wavelength of light, its shape, its refractive index, and the angle of observation. Historically, light scattering has been broadly categorized into two regimes: Rayleigh scattering for particles much smaller than the wavelength, and Mie scattering for particles comparable to or larger than the wavelength. While these two categories provide a useful starting point, a complete physical understanding requires a more sophisticated framework—one that treats light as an electromagnetic wave and accounts for diffraction, interference, and polarization. This is the domain of physical optics models.

Physical optics models go beyond simple geometric ray tracing by solving, or approximating, the wave equation under realistic boundary conditions. They are indispensable for situations where the wave nature of light is dominant, such as when scattering angles are large, when the particle is not spherical, or when multiple scattering events occur. In atmospheric physics, these models enable scientists to retrieve aerosol properties from satellite measurements, simulate the formation of halos and glories, and quantify the radiative forcing of clouds and pollutants.

Fundamentals of Light Scattering in the Atmosphere

The Role of Particle Size: Rayleigh versus Mie

The most basic distinction in atmospheric light scattering is provided by the ratio of the particle radius r to the wavelength λ. This dimensionless size parameter is defined as x = 2πr/λ. When x ≪ 1, the scattering falls into the Rayleigh regime, named after Lord Rayleigh who first derived the theory in the 19th century. In this regime, the scattering cross-section is proportional to λ−4, which is why shorter wavelengths (blue) scatter much more strongly than longer wavelengths (red). Rayleigh scattering is responsible for the blue sky and the reddening of sunlight at low angles.

When x is of order 1 or larger, the Rayleigh approximation fails and the full Mie theory must be used. Mie scattering, derived by Gustav Mie in 1908, applies to spherical particles of any size and provides exact solutions for the scattered field. Aerosols, water droplets in clouds, dust, and pollen grains typically fall into the Mie regime. Mie scattering is less wavelength-dependent than Rayleigh scattering, which explains why clouds appear white (all visible wavelengths are scattered similarly). It also produces strong forward scattering lobes and distinct angular patterns that are highly sensitive to particle size and composition.

Beyond Size: Shape, Composition, and Morphology

Real atmospheric particles are often non‑spherical—ice crystals are hexagonal prisms, mineral dust grains are irregular, and soot aggregates form complex fractal clusters. For such particles, Mie theory is insufficient because it assumes perfect sphericity. Physical optics models must account for the particle’s exact shape, often using numerical techniques such as the discrete dipole approximation (DDA), the finite‑difference time‑domain (FDTD) method, or the T‑matrix method. These methods solve Maxwell’s equations on a discretized representation of the particle volume or surface, allowing for arbitrary geometries.

Composition also matters. Particles may be homogeneous or layered, with coatings of different refractive indices. For instance, soot cores coated with organic matter or sulfate exhibit different scattering and absorption behavior than bare soot. Similarly, biological particles (pollen, bacteria) have refractive indices that vary within the particle. Accurate physical optics models can capture these internal structures and their effects on the phase matrix, which describes how the polarization state of light changes upon scattering.

Physical Optics Models: Theoretical Foundations

Maxwell’s Equations and the Scattering Problem

At the heart of any physical optics model are Maxwell’s equations. For a monochromatic plane wave incident on a particle embedded in a homogeneous medium, the total electric field can be written as the sum of the incident field and the scattered field. The goal is to find the scattered field at all points outside the particle. This constitutes a boundary‑value problem in which the fields inside and outside the particle must satisfy continuity conditions at the particle surface.

For spherical particles, the separation of variables in spherical coordinates yields the Mie solution as an infinite series of spherical harmonics and Bessel functions. For non‑spherical particles, analytical solutions are rarely available, so numerical methods are employed. Two widely used approaches are:

  • The T‑Matrix Method: This method expands the incident and scattered fields in vector spherical wave functions and relates them via a transition matrix (T‑matrix). It is highly efficient for axisymmetric particles (spheroids, cylinders) and can handle particles up to several tens of wavelengths in size.
  • The Discrete Dipole Approximation (DDA): The particle is represented as an array of polarizable dipoles. The response of each dipole to the local electric field is computed self‑consistently, yielding the scattering and absorption properties. DDA is flexible for arbitrary shapes but becomes computationally expensive for large particles.

The Born Approximation and Anomalous Diffraction

In some regimes, approximate analytical methods provide useful insight without full numerical computation. The Born approximation (first‑order perturbation theory) applies when the particle is optically soft (refractive index close to the surrounding medium). It treats the scattered field as a sum of contributions from every point inside the particle, assuming the incident field is only weakly perturbed. This works well for small refractive index contrasts, such as some biological particles or weakly absorbing aerosols.

The anomalous diffraction model, introduced by van de Hulst, is applicable for large, soft particles. It combines the refraction through the particle with diffraction around its edges, ignoring reflections. This model captures the broad forward scattering lobe and can be used to derive the total extinction cross‑section. It is particularly useful for cloud droplets and large pollen grains.

Key Wave Phenomena in Physical Optics Models

Diffraction

Diffraction arises from the wave nature of light and the finite extent of the particle. When a plane wave encounters a particle, the obstruction causes the wavefront to spread into the geometric shadow region. This is described by Fraunhofer diffraction for large particles (the familiar Airy pattern for a circular aperture). In physical optics scattering models, diffraction is automatically included in the full wave solution. For particles much larger than the wavelength, the diffraction pattern is strongly peaked in the forward direction, with an angular width inversely proportional to the particle size. This forward‑scatter peak is important in remote sensing: instruments that measure near‑forward scattering can infer particle size distributions from the diffraction pattern.

Interference

Interference occurs when two or more coherent waves overlap. In a scattering event, the scattered waves from different parts of the particle can interfere constructively or destructively, producing angular variations in intensity. For example, the rainbow is the result of constructive interference after one internal reflection inside a water droplet, an effect that requires a wave description (though the geometric optic ray path provides the primary angle). Physical optics models capture these interference fringes, which can be used to measure droplet size with high precision.

Interference also plays a role in the corona phenomenon—concentric colored rings around the sun or moon produced by diffraction and interference from small water droplets in thin clouds. The angular spacing of the rings is directly related to the droplet size distribution, making corona observations a low‑cost method for cloud microphysics monitoring.

Polarization

The polarization state of light—the orientation of the electric field vector—is profoundly modified by scattering. Rayleigh scattering produces completely polarized light at 90° scattering angle (the maximum polarization perpendicular to the scattering plane). Mie scattering from larger particles introduces complex polarization patterns that depend on particle size, shape, and orientation. Physical optics models compute the full Stokes vector of the scattered field, enabling accurate interpretation of polarimetric remote sensing data. For instance, space‑borne polarimeters such as the POLDER and MSPI instruments use multi‑angle, multi‑spectral polarization measurements to retrieve aerosol properties.

The degree of linear polarization (DoLP) is a sensitive indicator of particle shape: non‑spherical aerosols produce characteristically different angular polarization signatures than spherical ones. This is critical for distinguishing between mineral dust and sulfate aerosols in satellite retrievals.

Applications in Atmospheric Physics and Remote Sensing

The Blue Sky and Red Sunsets

The classic application of Rayleigh scattering is the blue sky. Sunlight passing through the atmosphere is Rayleigh‑scattered by air molecules (N₂ and O₂, which have sizes ~0.1 nm, far smaller than visible wavelengths). The scattering coefficient varies as λ⁻⁴, so blue light (450 nm) is scattered roughly 9 times more than red light (650 nm). At low sun angles, the path through the atmosphere is longer, and most short‑wavelength light is scattered away, leaving longer wavelengths to create the sunset’s red hues. Physical optics models that incorporate the full Rayleigh phase function and the molecular density profile can quantitatively reproduce sky radiance, as well as the observed polarization patterns.

Halos, Glories, and Sundogs

Ice crystals in cirrus clouds produce a rich variety of optical phenomena, including the 22° halo, sundogs, and circumzenithal arcs. These arise from refraction and reflection through hexagonal ice prisms. While geometric optics can predict the primary ray paths, physical optics models are necessary to explain the intensity distribution, chromatic structure, and polarization of halos. For example, the 22° halo’s sharp inner edge and diffuse outer edge are better captured by wave‑optical diffraction integral methods. Similarly, the glory—a set of colored rings observed from aircraft or mountains opposite the sun—results from backscattering from cloud droplets and involves resonant wave effects (surface waves and tunneling). Full Mie calculations reproduce the glory’s angular position and color sequence, but the phenomenon was only fully understood after the development of physical optics theory.

Aerosol Remote Sensing

Satellite‑based retrievals of aerosols rely on comparing measured top‑of‑atmosphere radiances with pre‑computed look‑up tables generated from physical optics scattering models. For example, the Moderate Resolution Imaging Spectroradiometer (MODIS) uses dark‑target, deep‑blue, and multi‑angle algorithms that assume aerosol models (e.g., spherical for sulfate, non‑spherical for dust) derived from T‑matrix or DDA calculations. Accurate modeling of the scattering phase function and single‑scattering albedo is crucial for distinguishing between absorbing (e.g., black carbon) and scattering aerosols (e.g., sulfates).

Ground‑based networks like AERONET also use sky radiance measurements at multiple angles and wavelengths to retrieve aerosol optical depth, size distribution, refractive index, and shape. The inversion codes used by AERONET are built on physical optics libraries such as Mätzler’s MATLAB Mie codes and the spheroid scattering model developed by Dubovik et al. These models must cover the full size range from submicron to tens of microns, often using a mixture of spheres and spheroids to represent real aerosols.

Cloud Microphysics and Radiative Transfer

Physical optics models are essential for understanding the radiative properties of clouds. Cloud droplets have sizes from a few microns to several tens of microns, falling squarely in the Mie regime. The asymmetry parameter g (the average cosine of the scattering angle) determines how much forward‑scattered radiation propagates through the cloud: for water clouds, g ≈ 0.85, meaning strong forward scattering. This influences the cloud albedo and the vertical distribution of heating rates. In climate models, parameterizations of cloud optical properties are derived from Mie calculations and then used in radiative transfer solvers like DISORT or the delta‑Eddington approximation.

Ice clouds are more challenging because ice crystals exist in many habits (columns, plates, needles, bullet rosettes) that scatter light differently. Physical optics models (e.g., the ray‑tracing with diffraction method or the improved geometric optics method) are used to compute the phase functions for these habits. The results feed into satellite retrieval algorithms (e.g., for the Cloud‑Aerosol Lidar with Orthogonal Polarization (CALIOP) on CALIPSO) and into climate models to account for the effects of ice clouds on the Earth’s radiation budget.

Advanced Topics and Emerging Techniques

Multiple Scattering in Turbid Media

When the optical thickness of an aerosol layer or cloud is large (τ > 0.1), multiple scattering becomes important. Each photon may undergo several scattering events before reaching the sensor or leaving the atmosphere. Physical optics models for multiple scattering solve the vector radiative transfer equation (VRTE), which accounts for the full Stokes vector and the phase matrix. Methods such as the Monte Carlo technique, the adding‑doubling method, and the matrix operator method are employed. These are critical for interpreting lidar returns (where multiple scattering reduces the effective extinction coefficient) and for high‑resolution imaging of land surfaces through hazy atmospheres.

Coherent Backscattering and Opposition Effect

A subtle but important multiple‑scattering effect is coherent backscattering (CBS), also known as the opposition effect. When a medium composed of many particles is illuminated, waves traversing the same path in opposite directions can interfere constructively, leading to a sharp peak in reflectivity exactly in the backscattering direction. This effect is observed in cloud decks, soil surfaces, and planetary regolith. Physical optics models that treat the full wave field in dense media (such as the radiative transfer with coherent backscattering theory) can reproduce the amplitude and angular width of this peak. In Earth remote sensing, the opposition effect is seen in satellite images of clouds and snow, and must be accounted for in albedo retrievals.

Super‑Resolution and Inverse Problems

Inverse problems aim to retrieve particle properties from measurements of scattered light. Because of the nonlinear mapping from particle size and shape to scattered intensity, physical optics models combined with optimization algorithms (e.g., Levenberg‑Marquardt, Bayesian inversion, neural networks) are used to solve the inverse problem. Recent advances include the use of deep learning surrogates for physical optics models, enabling real‑time retrieval of aerosol size distributions from lidar measurements. These hybrid approaches yield faster inversions while retaining the physical accuracy of the full wave models.

Conclusion and Outlook

Physical optics models provide an indispensable bridge between the fundamental wave theory of light and the complex reality of atmospheric particles. By explicitly including diffraction, interference, and polarization, these models capture the richness of scattering phenomena that simple geometric optics cannot predict. From explaining the colors of the sky to retrieving aerosol properties from space, physical optics underpins much of modern atmospheric science.

Future developments will likely focus on several areas: improved computational efficiency for ensemble‑averaged models of irregular particles (e.g., using precomputed databases or machine‑learning emulators), better characterization of complex morphologies such as fractal soot aggregates or biological particles, and integration with climate models to represent the rapid adjustments of aerosol‑cloud interactions. As observational techniques continue to advance—such as hyperspectral polarimetry, multi‑angle lidar, and high‑resolution imaging—physical optics models will remain essential for interpreting these data and deepening our understanding of Earth’s atmospheric system.

For further reading, see standard references: Rayleigh scattering on Wikipedia, Mie scattering, and the comprehensive textbook “Light Scattering by Particles: Computational Methods”. Additionally, the NASA GISS publications on light scattering provide extensive resources on T‑matrix and DDA applications in atmospheric physics.