Introduction
The boundary families summarized here correspond to the classical fluid, rigid, elastic, shell, and viscous-interface scattering literature (Anderson 1950; Faran 1951; Goodman and Stern 1962; Feuillade and Nero 1998).
Boundary conditions are the mathematical statements that turn a general wave equation into a specific scattering problem. In target-strength modeling, they specify what the target surface can and cannot do in response to the incident sound field. A boundary may be perfectly rigid, perfectly pressure releasing, fluid transmitting, weakly contrasting, or dynamically elastic. Those are not just different algebraic cases. They are different physical idealizations, and they produce different scattering mechanisms even for the same gross geometry.
This page gathers the main boundary families that recur throughout the theory articles. The aim is to explain both the mathematics and the physics clearly enough that a reader can understand what is being assumed before turning to any model-specific derivation.
It is best read alongside the shared acoustic
scattering primer and notation
guide. In the package theory pages, medium 1 is always
the surrounding seawater, medium 2 is the first target
region encountered from the outside, and deeper internal regions are
numbered inward.
The main boundary families separate most naturally according to the field quantity constrained most directly at the interface: pressure, normal velocity, layered acoustic transmission, or elastic traction.
Shared interface notation
Before specializing to any one interface, it helps to keep the common
two-medium notation explicit. If media i and j
meet at an interface, then the density and sound-speed contrasts
are:
g_{ij} = \frac{\rho_i}{\rho_j}, \qquad h_{ij} = \frac{c_i}{c_j},
and the acoustic wavenumbers are:
k_i = \frac{\omega}{c_i}, \qquad k_j = \frac{\omega}{c_j}.
For a one-region target in seawater, the most common contrasts are therefore g_{21} and h_{21}. For a shelled target, the shell-to-water interface uses g_{21} and h_{21}, while an internal fluid relative to the shell uses g_{32} and h_{32}.
Simple boundaries
Fixed rigid
A fixed rigid boundary is the limiting case of a scatterer whose surface cannot move in the normal direction. In an inviscid surrounding fluid, the acoustic pressure field p(\mathbf{x}, t) and the particle velocity field \mathbf{v}(\mathbf{x}, t) satisfy the linearized momentum equation:
\rho \frac{\partial \mathbf{v}}{\partial t} = -\nabla p,
where \rho is the surrounding-fluid density, \mathbf{x} is the position vector, and \nabla p is the pressure gradient. If \mathbf{n} denotes the outward unit normal vector to the target surface, then the normal component of particle velocity is:
v_n = \mathbf{v} \cdot \mathbf{n},
where the dot denotes the Euclidean inner product. The mechanical impedance at the boundary is then defined by:
Z_s = \frac{p}{v_n},
where Z_s is the surface impedance. For a rigid boundary, the surface cannot move, so:
v_n = 0.
With harmonic time dependence, that implies the Neumann boundary condition:
\frac{\partial p}{\partial n} = \nabla p \cdot \mathbf{n} = 0,
where \partial / \partial n denotes the derivative taken along the outward normal direction. The exterior acoustic field is therefore solved subject to:
\nabla^2 p + k^2 p = 0,
where k = \omega / c is the surrounding-fluid wavenumber, \omega is angular frequency, and c is sound speed. Physically, the interface stores no normal motion. In partial-wave and modal formulations, that condition fixes the scattering coefficients by eliminating any solution that would produce radial displacement of the boundary. In the high-frequency limit, a sufficiently smooth rigid surface supports specular reflection without the phase inversion associated with a pressure-release boundary.
Pressure release
A pressure-release boundary is the opposite limiting case. Here the surface cannot sustain an acoustic pressure fluctuation, so the interface behaves like a perfectly soft boundary. In impedance language:
Z_s = \frac{p}{v_n} \to 0.
For finite normal motion, the only consistent condition is:
p = 0.
This is a Dirichlet boundary condition. It removes any admissible solution of the Helmholtz equation that carries nonzero pressure at the interface. Physically, this corresponds to a surface that relieves pressure fluctuations essentially instantaneously, as in an idealized gas interface or other highly compliant boundary. Relative to the rigid case, the reflected pressure undergoes a phase reversal. In modal descriptions, the scattering coefficients are fixed by enforcing zero pressure rather than zero normal velocity at the surface.
Fluid-filled
For a fluid-filled scatterer, both the exterior and interior regions support acoustic pressure fields. Let medium 1 denote the surrounding fluid and medium 2 the interior fluid. Then:
\nabla^2 p_1 + k_1^2 p_1 = 0, \qquad \nabla^2 p_2 + k_2^2 p_2 = 0,
where p_1 and p_2 are the exterior and interior pressures, and:
k_j = \frac{\omega}{c_j}, \qquad j = 1, 2,
with c_j the sound speed in medium j. At the interface, continuity of normal traction gives continuity of acoustic pressure:
p_1 = p_2.
Continuity of normal particle velocity gives:
\mathbf{v}_1 \cdot \mathbf{n} = \mathbf{v}_2 \cdot \mathbf{n},
where \mathbf{v}_1 and \mathbf{v}_2 are the particle-velocity fields in the two media. Using the linearized momentum relation in each fluid:
\mathbf{v}_j = -\frac{1}{i \omega \rho_j}\nabla p_j,
the normal-velocity condition can also be written as:
\frac{1}{\rho_1}\frac{\partial p_1}{\partial n} = \frac{1}{\rho_2}\frac{\partial p_2}{\partial n}.
These are the standard transmission conditions for an acoustic fluid-fluid boundary (Morse and Ingard 1968; Medwin and Clay 1998). Neither pressure nor normal motion is forced to vanish. Instead, both media participate in the interface dynamics, and the scattering is controlled by density and compressibility contrast across the boundary.
Inelastic shelled boundaries
The inelastic shelled cases are layered acoustic idealizations rather than true elastic shell problems. The shell is treated as a non-deforming, non-shear-supporting layer that constrains the adjacent acoustic fields kinematically. That makes these boundaries useful as simplified layered models, but they should be distinguished clearly from a genuinely elastic shell that supports both compressional and shear waves.
Pressure release interior
In this configuration, an inelastic shell surrounds a pressure-release interior. The shell is assumed not to support elastic wave propagation and not to deform according to Hooke’s law. It therefore acts as a kinematic constraint rather than as a dynamic elastic medium. A convenient way to describe the surrounding-fluid and shell-adjacent acoustic fields is to introduce an exterior pressure field p_{\text{exterior}} and an acoustic field immediately adjacent to the shell, written here as p_{\text{shell}}.
At the shell-exterior interface, continuity of pressure and normal velocity with the surrounding fluid is enforced:
p_{\text{shell}} = p_{\text{exterior}}, \qquad v_{n,\text{shell}} = v_{n,\text{exterior}},
where v_{n,\text{shell}} and v_{n,\text{exterior}} are the normal particle velocities just inside and outside the shell interface, respectively. At the shell-interior interface, the interior is pressure releasing, so:
p_{\text{shell}} = 0, \qquad v_{n,\text{shell}} \neq 0.
This idealization means that the shell must transmit normal motion while accommodating a zero-pressure cavity on its inner side. The shell itself does not carry elastic shell-wave physics, but it does modify how the exterior acoustic field couples to the interior cavity. In scattering formulations, the resulting layered boundary conditions differ from both rigid and ordinary fluid-filled cases and can introduce low-frequency cavity-type resonances.
Mathematically, this case may be viewed as an acoustic transmission problem with an outer continuity condition and an inner Dirichlet condition. The shell is therefore acting as a geometric and kinematic intermediary between the exterior fluid and a pressure-release cavity, not as an elastic medium in its own right.
Fluid-filled interior
For an inelastic shell enclosing a fluid-filled interior, the shell is taken to have effectively infinite mechanical impedance and therefore negligible normal displacement:
v_{n,\text{shell}} = 0.
At the shell-exterior interface, that produces a Neumann condition on the acoustic field:
\frac{\partial p_{\text{shell}}}{\partial n} = 0.
At the shell-interior fluid interface, the interior fluid must satisfy continuity of pressure and normal velocity with the shell-adjacent field:
p_{\text{interior}} = p_{\text{shell}}, \qquad v_{n,\text{interior}} = v_{n,\text{shell}},
where p_{\text{interior}} is the interior-fluid pressure and v_{n,\text{interior}} is the interior-fluid normal velocity at the inner shell surface. Because the shell is inelastic, one typically also takes:
v_{n,\text{shell}} \approx 0.
The interior fluid therefore experiences an effectively rigid boundary while still supporting compressional waves of its own. In that sense, the system is a layered acoustic cavity rather than an elastic shell. The resulting scattering is controlled by acoustic layering and cavity behavior rather than by shell-wave propagation.
To make that structure explicit, the interior acoustic field still satisfies:
\nabla^2 p_{\text{interior}} + k_{\text{interior}}^2 p_{\text{interior}} = 0,
The associated interior wavenumber is:
k_{\text{interior}} = \frac{\omega}{c_{\text{interior}}},
where c_{\text{interior}} is the sound speed of the interior fluid. The shell therefore modifies the interior field through boundary constraints even though it does not itself support elastic motion.
Elastic shelled boundaries
Elastic shell boundary
An elastic shell supports both longitudinal and transverse elastic motion (Achenbach 1973). The displacement field \mathbf{u}(\mathbf{x}, t) within the shell satisfies the Navier equation:
(\lambda + 2 \mu)\nabla(\nabla \cdot \mathbf{u}) - \mu \nabla \times (\nabla \times \mathbf{u}) + \rho_s \frac{\partial^2 \mathbf{u}}{\partial t^2} = 0,
where \lambda and \mu are the Lamé parameters and \rho_s is the shell density. Using the Helmholtz decomposition:
\mathbf{u} = \nabla \phi + \nabla \times \mathbf{\Psi},
the displacement separates into longitudinal and transverse potentials, each satisfying its own Helmholtz equation with wavenumbers:
k_\ell = \frac{\omega}{c_\ell}, \qquad k_\tau = \frac{\omega}{c_\tau},
and wave speeds:
c_\ell = \sqrt{\frac{\lambda + 2 \mu}{\rho_s}}, \qquad c_\tau = \sqrt{\frac{\mu}{\rho_s}}.
The elastic stress tensor is:
\sigma_{ij} = \lambda \delta_{ij}\nabla \cdot \mathbf{u} + 2 \mu \varepsilon_{ij},
where \delta_{ij} is the Kronecker delta and \varepsilon_{ij} is the infinitesimal strain tensor:
\varepsilon_{ij} = \frac{1}{2} \left( \frac{\partial u_i}{\partial x_j} + \frac{\partial u_j}{\partial x_i} \right).
At a fluid-solid interface, the fluid pressure must balance the normal traction in the shell, the normal fluid velocity must match the normal shell velocity, and the tangential traction must vanish because the fluid is inviscid. If u_n = \mathbf{u} \cdot \mathbf{n} is the shell displacement in the normal direction, \sigma_{nn} is the normal traction, and \sigma_{nt} is the tangential traction, then the interface conditions may be written schematically as:
p = -\sigma_{nn}, \qquad -\frac{1}{i \omega \rho_f}\frac{\partial p}{\partial n} = -i \omega u_n, \qquad \sigma_{nt} = 0,
where \rho_f is the fluid density. The exact sign of the pressure-traction relation depends on the outward-normal convention, but the physical content is unchanged. The shell carries both normal and tangential motion, and that coupling makes the scattering problem much richer than either the rigid or the inelastic-shell cases. The resulting modal systems support resonance structure associated with shell thickness, elastic moduli, and both the exterior and interior fluid properties.
Weak, fluid-like
The weak fluid-like case is best understood as an approximation regime rather than as a sharply enforcing boundary type. The main assumption is that the target differs only slightly from the surrounding medium, so the acoustic field inside the scatterer remains close to the incident field. In that limit:
p_{\text{interior}} \approx p_{\text{exterior}}.
The corresponding material contrasts are small. If \rho_1 and c_1 denote the surrounding-fluid density and sound speed, and \rho_2 and c_2 denote the target values, one typically writes:
g = \frac{\rho_2}{\rho_1} \approx 1, \qquad h = \frac{c_2}{c_1} \approx 1.
Equivalent density and compressibility contrasts may be written as:
\gamma_\rho = \frac{\rho_2 - \rho_1}{\rho_2}, \qquad \gamma_\kappa = \frac{\kappa_2 - \kappa_1}{\kappa_1},
where \kappa_j = (\rho_j c_j^2)^{-1} is compressibility in medium j (Medwin and Clay 1998). The scattered field is then treated as a first-order perturbation produced by these small departures from the surrounding fluid rather than by a strongly reflecting or traction-supporting surface.
This is why weak-scattering treatments are usually expressed as perturbative approximations instead of exact boundary-value problems with a sharply enforcing interface condition. Physically, the target behaves approximately like the surrounding fluid, and the backscatter is produced by residual density and compressibility contrasts rather than by a large impedance jump at the boundary.