Skip to content

Helmholtz Theorem and Seismic Wave Propagation

CSIR NET Earth Sciences Part C: Helmholtz decomposition, Navier equation, P and S wave equations, elastic moduli, plane waves, reflection, surface waves, Q.

  • 17 min read
  • Updated
On this page · 15 sections
  1. Prerequisites
  2. The Hook: Why Two Body Waves?
  3. The Helmholtz Theorem
  4. From Newton’s Law to the Wave Equations
  5. Plane and Spherical Waves
  6. Waves at Interfaces
  7. Surface Waves
  8. Attenuation and Q
  9. Memory Tricks
  10. Don’t Confuse
  11. Common Mistakes
  12. Practice
  13. Quick Revision Table
  14. Key Takeaways
  15. What’s Next

Prerequisites

The Hook: Why Two Body Waves?

The Helmholtz Theorem

Statement

The proof uses only \( \nabla^2\left(\frac{1}{4\pi\lvert\mathbf{r}-\mathbf{r}'\rvert}\right) = -\delta(\mathbf{r}-\mathbf{r}') \), the same Green’s function as the Newtonian potential. Taking the divergence of the decomposition recovers \( \nabla\cdot\mathbf{F} \) from \( -\nabla^2\Phi \); taking the curl recovers \( \nabla\times\mathbf{F} \) (using \( \nabla\cdot\mathbf{A} = 0 \)).

PartPropertySource
\( -\nabla\Phi \)curl-free (irrotational, longitudinal)the divergence of \( \mathbf{F} \)
\( \nabla\times\mathbf{A} \)divergence-free (solenoidal, transverse)the curl of \( \mathbf{F} \)

The theorem across geophysics

FieldDivergenceCurlHelmholtz result
Gravity \( \mathbf{g} \)\( -4\pi G\rho \)0\( \mathbf{g} = \nabla U \), \( U = G\int\rho/r\,dV \): the Newtonian potential
Static magnetic \( \mathbf{B} \)0\( \mu_0\mathbf{J} \)\( \mathbf{B} = \nabla\times\mathbf{A} \), \( \mathbf{A} = \frac{\mu_0}{4\pi}\int\mathbf{J}/r\,dV \) (Biot–Savart)
Electrostatic \( \mathbf{E} \)\( \rho_q/\varepsilon_0 \)0\( \mathbf{E} = -\nabla\phi \), Coulomb potential
Seismic displacement \( \mathbf{u} \)dilatation \( \theta \)rotationP part plus S part (below)
Ocean or atmospheric velocityhorizontal divergencevorticityvelocity potential plus stream function

So gravity, magnetostatics and elastic waves are three applications of one theorem.

From Newton’s Law to the Wave Equations

Equation of motion and Hooke’s law

For an elastic solid of density \( \rho \), Newton’s second law per unit volume (ignoring gravity and body forces) is

$$\rho\,\frac{\partial^2 u_i}{\partial t^2} = \frac{\partial\tau_{ij}}{\partial x_j}$$

For an isotropic, linearly elastic solid, Hooke’s law with the Lamé parameters \( \lambda \) and \( \mu \) (the shear modulus) is

$$\tau_{ij} = \lambda\,\theta\,\delta_{ij} + 2\mu\,e_{ij}, \qquad \theta = \nabla\cdot\mathbf{u}, \qquad e_{ij} = \tfrac{1}{2}\left(\partial_i u_j + \partial_j u_i\right)$$

Substituting, for a homogeneous medium:

$$\boxed{\rho\,\ddot{\mathbf{u}} = (\lambda + 2\mu)\,\nabla(\nabla\cdot\mathbf{u}) - \mu\,\nabla\times(\nabla\times\mathbf{u})}$$

This is the Navier equation (equivalently \( \rho\ddot{\mathbf{u}} = (\lambda+\mu)\nabla(\nabla\cdot\mathbf{u}) + \mu\nabla^2\mathbf{u} \)).

Separating P and S

Write the displacement with Helmholtz potentials:

$$\boxed{\mathbf{u} = \nabla\phi + \nabla\times\boldsymbol{\psi}, \qquad \nabla\cdot\boldsymbol{\psi} = 0}$$

Then \( \nabla\cdot\mathbf{u} = \nabla^2\phi \) and \( \nabla\times\nabla\times\mathbf{u} = -\nabla\times\nabla^2\boldsymbol{\psi} \). The Navier equation becomes

$$\nabla\left[(\lambda+2\mu)\nabla^2\phi - \rho\ddot{\phi}\right] + \nabla\times\left[\mu\nabla^2\boldsymbol{\psi} - \rho\ddot{\boldsymbol{\psi}}\right] = 0$$

which is satisfied when each bracket vanishes:

$$\boxed{\nabla^2\phi = \frac{1}{\alpha^2}\frac{\partial^2\phi}{\partial t^2}, \quad \alpha = \sqrt{\frac{\lambda+2\mu}{\rho}} = \sqrt{\frac{K + \frac{4}{3}\mu}{\rho}}}$$$$\boxed{\nabla^2\boldsymbol{\psi} = \frac{1}{\beta^2}\frac{\partial^2\boldsymbol{\psi}}{\partial t^2}, \quad \beta = \sqrt{\frac{\mu}{\rho}}}$$

Equivalently, take the divergence and the curl of the Navier equation directly: the dilatation \( \theta = \nabla\cdot\mathbf{u} \) travels at \( \alpha \), and the rotation \( \nabla\times\mathbf{u} \) travels at \( \beta \).

P (primary, compressional)S (secondary, shear)
Potentialscalar \( \phi \)vector \( \boldsymbol{\psi} \)
Divergence of \( \mathbf{u} \)non-zero (volume change)zero
Curl of \( \mathbf{u} \)zero (no rotation)non-zero
Particle motionparallel to propagationperpendicular to propagation (SV in the vertical plane, SH horizontal)
Speed\( \alpha = \sqrt{(K + 4\mu/3)/\rho} \)\( \beta = \sqrt{\mu/\rho} \)
In a fluid (\( \mu = 0 \))\( \alpha = \sqrt{K/\rho} \)does not exist

Consequences

  • P is always faster than S. \( \alpha^2/\beta^2 = (K/\mu) + 4/3 > 4/3 \) for any material with positive bulk modulus.
  • No S waves in fluids. The outer core, oceans and magma do not transmit S; this is the evidence for a liquid outer core.
  • Poisson solid. With \( \lambda = \mu \) (a good first approximation for many crustal rocks), \( \alpha = \sqrt{3}\,\beta \).
  • Moduli from velocities: \( \mu = \rho\beta^2 \), \( K = \rho\left(\alpha^2 - \frac{4}{3}\beta^2\right) \).

Poisson’s ratio from \( V_P/V_S \)

$$\boxed{\nu = \frac{\alpha^2 - 2\beta^2}{2\left(\alpha^2 - \beta^2\right)}}$$
\( V_P/V_S \)\( \nu \)Typical meaning
1.50.10stiff, low-\( \nu \) material
\( \sqrt{3} \approx 1.73 \)0.25Poisson solid
2.00.33high \( \nu \)
\( \to\infty \)\( \to 0.5 \)fluid (\( \beta = 0 \))

High \( V_P/V_S \) is often read as a sign of fluid or partial melt; the ratio is used in crustal and reservoir studies for that reason.

Plane and Spherical Waves

Plane waves

A harmonic plane wave has potentials \( \phi = \Phi_0\, e^{i(\mathbf{k}\cdot\mathbf{x} - \omega t)} \) with \( \lvert\mathbf{k}\rvert = \omega/\alpha \). Then

$$\mathbf{u} = \nabla\phi = i\mathbf{k}\,\phi \quad\Rightarrow\quad \mathbf{u}\parallel\mathbf{k} \quad \text{(longitudinal)}$$

For the S potential, \( \mathbf{u} = \nabla\times\boldsymbol{\psi} = i\mathbf{k}\times\boldsymbol{\psi} \), so \( \mathbf{u}\perp\mathbf{k} \) (transverse), with \( \lvert\mathbf{k}\rvert = \omega/\beta \).

Quick test of a displacement field. Compute \( \nabla\cdot\mathbf{u} \) and \( \nabla\times\mathbf{u} \):

\( \mathbf{u} \)\( \nabla\cdot\mathbf{u} \)\( \nabla\times\mathbf{u} \)Wave
\( (A\sin k(x - ct),\ 0,\ 0) \)\( Ak\cos k(x-ct) \neq 0 \)0P travelling along \( x \)
\( (0,\ A\sin k(x - ct),\ 0) \)0\( (0,0,Ak\cos k(x-ct)) \neq 0 \)S travelling along \( x \), polarised along \( y \)
\( (0,\ 0,\ A\sin k(x - ct)) \)0non-zeroS along \( x \), polarised along \( z \)

With \( z \) vertical and propagation along horizontal \( x \), the \( y \)-polarised wave is SH and the \( z \)-polarised one is SV.

Spherical waves and geometrical spreading

From a point source, the P potential is \( \phi = \frac{1}{r}f\!\left(t - \frac{r}{\alpha}\right) \): the amplitude falls as \( 1/r \) because the energy spreads over a sphere of area \( 4\pi r^2 \).

WaveSpreads overAmplitude falls asEnergy falls as
Body wavesphere, \( \propto r^2 \)\( 1/r \)\( 1/r^2 \)
Surface wavecylinder (ring), \( \propto r \)\( 1/\sqrt{r} \)\( 1/r \)

That is why surface waves dominate seismograms of shallow earthquakes at large distances.

Waves at Interfaces

Snell’s law and the ray parameter

At a plane interface the horizontal slowness is conserved:

$$\boxed{p = \frac{\sin i_1}{v_1} = \frac{\sin i_2}{v_2}}$$

for every reflected and transmitted wave, including converted waves (P to SV and SV to P). SH does not convert to P or SV at a horizontal interface; P–SV and SH are decoupled. When \( v_2 > v_1 \), the critical angle is \( \sin i_c = v_1/v_2 \), beyond which the head wave used in refraction surveys is generated.

Normal incidence

At normal incidence there is no mode conversion. With acoustic impedance \( Z = \rho v \), the usual reflection-seismology coefficients are for stress (pressure) amplitude:

$$\boxed{R = \frac{Z_2 - Z_1}{Z_2 + Z_1}}, \qquad T = \frac{2Z_2}{Z_1 + Z_2} \quad\text{(stress amplitudes)}$$

For displacement (or particle velocity) the signs and transmission factor differ: \( R_u = (Z_1 - Z_2)/(Z_1 + Z_2) = -R \) and \( T_u = 2Z_1/(Z_1 + Z_2) \). Check which quantity a question means.

The fractions of incident energy reflected and transmitted are \( R^2 \) and \( \frac{Z_1}{Z_2}T^2 = 1 - R^2 \), the same in either convention. A negative \( R \) means the reflected stress wave has reversed polarity (going into lower impedance). The transmitted stress amplitude can exceed the incident one when \( Z_2 > Z_1 \), and the transmitted displacement amplitude can exceed the incident one when \( Z_2 < Z_1 \); energy is still conserved.

Surface Waves

A free surface (zero traction) or a layered medium allows waves trapped near the boundary.

RayleighLove
Built frominterfering P and SVtrapped SH
Needsa free surface (exists even on a uniform half-space)a low-velocity layer over a faster half-space (\( \beta_1 < \beta_2 \))
Particle motionelliptical in the vertical plane; retrograde at the surface of a half-space, reversing with depthhorizontal, transverse to propagation
Speedon a Poisson half-space, \( c_R = 0.919\,\beta \)between \( \beta_1 \) and \( \beta_2 \)
Dispersionnone on a uniform half-space; dispersive in layered mediaalways dispersive
Recorded onvertical and radial componentstransverse component

For a Poisson solid the Rayleigh equation gives \( c_R^2/\beta^2 = 2 - 2/\sqrt{3} \), so \( c_R = 0.9194\,\beta \). For Love waves, long wavelengths sample the faster half-space and travel near \( \beta_2 \); short wavelengths are confined to the layer and travel near \( \beta_1 \).

Phase and group velocity. In a dispersive wave train, crests move at the phase velocity \( c = \omega/k \), but energy moves at the group velocity

$$U = \frac{d\omega}{dk} = c + k\frac{dc}{dk}$$

Dispersion curves of surface waves (velocity against period) are inverted for the shear-velocity structure of the crust and upper mantle: longer periods sample deeper.

Interface waves also exist at a solid–solid boundary (Stoneley waves) and at a solid–fluid boundary such as the sea floor (Scholte waves). In borehole logging, the guided “Stoneley” or tube wave travels along the wall of the fluid-filled hole.

Attenuation and Q

Real rocks lose energy to internal friction. The quality factor \( Q \) is defined by the fractional energy loss per cycle, \( \frac{2\pi}{Q} = \frac{\Delta E}{E} \). For a wave travelling for time \( t \):

$$\boxed{A = A_0\, e^{-\pi f t/Q}} = A_0\, e^{-\pi f r/(Qv)}$$
  • Higher frequencies are lost faster (more cycles in the same distance): the Earth is a low-pass filter, which is why distant events sound “dull” and why deep seismic reflection uses lower frequencies.
  • Low Q means strong attenuation. Partially molten, hot or fluid-filled zones have low Q.
  • Observed amplitude = source × geometrical spreading × intrinsic attenuation × scattering × site effects.

Memory Tricks

Don’t Confuse

ThisvsThat
Irrotational part \( \nabla\phi \): P, volume change, no rotationsolenoidal part \( \nabla\times\boldsymbol{\psi} \): S, rotation, no volume change
SV: polarised in the vertical plane, couples with PSH: horizontal polarisation, decoupled, makes Love waves
Rayleigh: P–SV, vertical-plane ellipseLove: SH, horizontal transverse
Phase velocity \( \omega/k \): crestsgroup velocity \( d\omega/dk \): energy
Amplitude coefficient \( R \) (can be negative)energy fraction \( R^2 \)
Geometrical spreading (no energy lost)intrinsic attenuation \( Q \) (energy turned into heat)
Bulk modulus \( K \) (volume)shear modulus \( \mu \) (shape); fluids have \( \mu = 0 \)

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper.

Level 1: Foundation

  1. Which part of the Helmholtz decomposition of seismic displacement carries the P wave?
  2. What is Poisson’s ratio for a Poisson solid?
  3. A wave goes from a layer with \( v_1 = 2000 \) m/s into one with \( v_2 = 4000 \) m/s. What is the critical angle?
Answers
  1. The irrotational (gradient) part \( \nabla\phi \).
  2. 0.25 (\( \lambda = \mu \), \( V_P/V_S = \sqrt{3} \)).
  3. 30°. \( \sin i_c = 2000/4000 = 0.5 \).

Level 2: Part B Style

Practice Q1. A rock has \( \lambda = \mu = 30 \) GPa and density 2700 kg/m³. Its P-wave speed is about:

(a) 3.33 km/s (b) 4.71 km/s (c) 5.77 km/s (d) 6.67 km/s

Answer: (c) 5.77 km/s

\( \alpha = \sqrt{(30 + 60)\times 10^9/2700} = \sqrt{3.333\times 10^7} = 5774 \) m/s.

(a) is \( \beta = \sqrt{30\times 10^9/2700} = 3333 \) m/s. (b) uses \( \sqrt{(\lambda+\mu)/\rho} \).

Practice Q2. A sandstone has \( V_P/V_S = 2.0 \). Its Poisson’s ratio is:

(a) 0.25 (b) 0.30 (c) 0.33 (d) 0.40

Answer: (c) 0.33

\( \nu = \dfrac{4 - 2}{2(4 - 1)} = \dfrac{2}{6} = 0.333 \).

Practice Q3. The displacement field \( \mathbf{u} = (0,\ A\sin k(x - ct),\ 0) \), with \( z \) vertical, describes:

(a) a P wave travelling along \( y \) (b) an SH wave travelling along \( x \) (c) an SV wave travelling along \( x \) (d) a Rayleigh wave

Answer: (b)

It depends on \( x - ct \), so it travels along \( x \). \( \nabla\cdot\mathbf{u} = \partial u_y/\partial y = 0 \), \( (\nabla\times\mathbf{u})_z = \partial u_y/\partial x \neq 0 \): transverse, with horizontal (\( y \)) polarisation: SH.

Practice Q4. Taking the stress-amplitude convention, a P wave travels at normal incidence from shale (2400 kg/m³, 3000 m/s) into sandstone (2200 kg/m³, 2500 m/s). The reflection coefficient is:

(a) +0.134 (b) −0.134 (c) −0.018 (d) −0.091

Answer: (b) −0.134

\( R = (5.5 - 7.2)/(5.5 + 7.2) = -0.134 \). (c) is \( -R^2 \), the energy fraction with a sign. (d) uses velocities only: \( (2500 - 3000)/(5500) \).

Level 3: Part C Style

Practice Q5. A crustal rock has \( V_P = 6.0 \) km/s, \( V_S = 3.464 \) km/s and density 2700 kg/m³. Which statements are correct?

  1. Its shear modulus is about 32.4 GPa.
  2. Its bulk modulus is about 54 GPa.
  3. Its Poisson’s ratio is 0.25.
  4. Rayleigh waves on a half-space of this rock travel at about 3.18 km/s.

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 3 and 4 only (d) 1, 2, 3 and 4

Answer: (d) all four

\( \beta^2 = (3464)^2 = 1.200\times 10^7 \) m²/s², \( \alpha^2 = 3.6\times 10^7 \) m²/s².

  1. \( \mu = \rho\beta^2 = 2700\times 1.2\times 10^7 = 3.24\times 10^{10} \) Pa. Correct.
  2. \( K = \rho(\alpha^2 - \tfrac{4}{3}\beta^2) = 2700(3.6 - 1.6)\times 10^7 = 5.4\times 10^{10} \) Pa. Correct.
  3. \( \alpha^2/\beta^2 = 3 \): \( \nu = (3 - 2)/(2\times 2) = 0.25 \). Correct.
  4. Poisson solid, so \( c_R = 0.9194\times 3.464 = 3.18 \) km/s. Correct.

Practice Q6. For the shale-to-sandstone interface of Q4 (normal incidence), which statements are correct?

  1. About 1.8% of the incident energy is reflected.
  2. The transmitted stress amplitude is about 0.87 times the incident amplitude.
  3. The transmitted displacement amplitude is about 1.13 times the incident amplitude, which violates energy conservation.
  4. The reflected stress wave has reversed polarity.

(a) 1 and 4 only (b) 1, 2 and 4 only (c) 2 and 3 only (d) 1, 2, 3 and 4

Answer: (b) 1, 2 and 4 only
  1. \( R^2 = 0.134^2 = 0.018 \). Correct.
  2. \( T = 2Z_2/(Z_1 + Z_2) = 11.0/12.7 = 0.866 \). Correct.
  3. The displacement transmission factor \( T_u = 2Z_1/(Z_1 + Z_2) = 14.4/12.7 = 1.134 \) is indeed above 1, but the transmitted energy fraction \( = (Z_1/Z_2)T^2 = (7.2/5.5)(0.750) = 0.982 = 1 - R^2 \). Energy is conserved. Wrong.
  4. \( R < 0 \). Correct.

Practice Q7. A 10 Hz P wave travels for 2 s through rock with \( Q = 100 \). Which statements are correct?

  1. Intrinsic attenuation alone reduces its amplitude to about 53% of the initial value.
  2. A 20 Hz wave on the same path would retain about 28%.
  3. Between 100 km and 400 km from a shallow source, geometrical spreading reduces surface-wave amplitude by half and body-wave amplitude to a quarter.
  4. Raising Q to 200 would increase the loss.

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 3 and 4 only (d) 1, 3 and 4 only

Answer: (b) 1, 2 and 3 only
  1. \( e^{-\pi(10)(2)/100} = e^{-0.628} = 0.534 \). Correct.
  2. \( e^{-\pi(20)(2)/100} = e^{-1.257} = 0.285 \). Correct.
  3. Distance ×4: surface waves \( 1/\sqrt{4} = 1/2 \); body waves \( 1/4 \). Correct.
  4. Higher Q means less attenuation. Wrong.

Quick Revision Table

IdeaFormula or fact
Helmholtz\( \mathbf{F} = -\nabla\Phi + \nabla\times\mathbf{A} \); fixed by div and curl
Navier\( \rho\ddot{\mathbf{u}} = (\lambda+2\mu)\nabla(\nabla\cdot\mathbf{u}) - \mu\nabla\times\nabla\times\mathbf{u} \)
Potentials\( \mathbf{u} = \nabla\phi + \nabla\times\boldsymbol{\psi} \), \( \nabla\cdot\boldsymbol{\psi} = 0 \)
P speed\( \alpha = \sqrt{(\lambda+2\mu)/\rho} = \sqrt{(K + 4\mu/3)/\rho} \)
S speed\( \beta = \sqrt{\mu/\rho} \); zero in fluids
Poisson’s ratio\( \nu = (\alpha^2 - 2\beta^2)/[2(\alpha^2 - \beta^2)] \); \( \sqrt{3} \to 0.25 \)
Snell\( p = \sin i/v \) conserved; \( \sin i_c = v_1/v_2 \)
Normal incidence\( R = (Z_2 - Z_1)/(Z_2 + Z_1) \) (stress amplitude); energy \( R^2 \)
RayleighP–SV, retrograde at surface, \( 0.919\beta \) (Poisson half-space)
LoveSH, needs low-velocity layer, \( \beta_1 < c < \beta_2 \), dispersive
Spreadingbody \( 1/r \), surface \( 1/\sqrt{r} \)
Attenuation\( A = A_0 e^{-\pi f t/Q} \)

Key Takeaways

What’s Next

Advertisement
Advertisement