
Figure 4.1. P-wave: longitudinal (compressional) particle motion — particles move parallel to the ray. Python-generated — assets/scripts/fig_pwave_swave_motion.py
S = Secondary · Shear · Transverse
Particle motion perpendicular to propagation — exists in solids only
Two independent polarizations:
| Polarization | Plane of motion | Mode conversion at interface? |
|---|---|---|
| SV | Vertical plane of the ray | Yes → converts to P or Rayleigh |
| SH | Horizontal, ⊥ to ray plane | No → generates Love waves only |

Figure 4.2. S-wave: transverse (shear) particle motion — particles move perpendicular to the ray. Python-generated — assets/scripts/fig_pwave_swave_motion.py

Figure 4.3. SV motion lies in the vertical plane of the ray; SH motion is horizontal and perpendicular to it. Python-generated — assets/scripts/fig_sv_sh_polarization.py
In a fluid: — but the formula is not the reason.
The physical argument:
Consequences in this course:
Free surface boundary (zero traction) allows guided waves that decay as and are dispersive:
| Type | Particle motion | Speed | Requires |
|---|---|---|---|
| Rayleigh | Retrograde ellipse (P + SV) | Any elastic half-space | |
| Love | Horizontal SH only | Velocity layering |
Both are slower than body waves and carry the largest amplitudes at teleseismic distances.

Figure 4.4. Rayleigh: retrograde elliptical decay with depth (left, center). Love: SH trapped in slow surface layer by total internal reflection (right). Python-generated — assets/scripts/fig_surface_waves.py
Rayleigh waves exist in any elastic half-space — they are a natural free-surface solution.
Love waves require a slow layer over a faster half-space ():
A homogeneous half-space has Rayleigh but not Love waves — observing Love waves requires a layered Earth.
For those interested in planetary science: NASA's InSight used surface wave dispersion from marsquakes to map the Martian crustal layering.

Figure 4.5. spans ~100× from dry clay (60 m/s) to steel (~6000 m/s). Soft sediments can be 50× slower than basement rock. Python-generated — assets/scripts/fig_seismic_velocities.py
| Material state | ||
|---|---|---|
| Typical crustal rock | 0.25 | |
| Dry, cracked rock | 0.10–0.20 | 1.45–1.60 |
| Water-saturated sediment | 0.45–0.49 | 3.0–10.0 |
| Perfect fluid | 0.50 |
High → fluid saturation, magma, high pore pressure
Low → gas sand, dry fractured rock
Seattle example: Duwamish Valley (water-saturated alluvium) — why Pioneer Square shakes harder than Capitol Hill.
P and S travel the same distance at speeds :
One seismometer + one clock = earthquake distance
Used in real time by PNSN and ShakeAlert
s, s, km/s,
Seattle → Portland ≈ 280 km — consistent with a Cascades or Willamette Valley source.
| Component | Most sensitive to |
|---|---|
| Vertical (Z) | P-wave (compressional, vertical motion); Rayleigh wave (vertical ellipse component) |
| Horizontal N–S, E–W | S-wave (transverse); Love wave (horizontal SH); Rayleigh wave (horizontal ellipse component) |
A Love wave has no vertical component — a vertical-only seismometer misses it entirely.
This is why three-component instruments are essential for full wave-type identification.
The USGS ShakeAlert system detects fast-arriving P-waves to issue alerts before more damaging S-waves arrive.
For a Cascadia M9:
That 60–90 s warning window = time to stop trains, pause surgeries, move away from windows.
The physics: — always.
= time-averaged shear velocity in the top 30 m of soil
| Site Class | (m/s) | Description |
|---|---|---|
| A | > 1500 | Hard rock |
| B | 760–1500 | Rock |
| C | 360–760 | Dense soil / soft rock |
| D | 180–360 | Stiff soil |
| E | < 180 | Soft soil |
Design earthquake force for Class E is 3–5× larger than Class B.
In Seattle: Capitol Hill (glacial till, ~500 m/s) vs. Pioneer Square (artificial fill, ~180 m/s).
Deep learning models (PhaseNet, EQTransformer) pick P and S arrivals because of the physics from this lecture.
In-class prompt — try this now:
"A seismometer's vertical channel shows a sharp onset at 32 s; horizontal channels show a larger onset at 57 s. What wave types are these, and what can I estimate from the 25-second difference?"
Evaluate the AI response:
AI passes the formula test easily. The harder test is physical reasoning — not algebra.
A seismometer records only P-wave arrivals — no S-wave. List three distinct physical reasons this could happen. (Think about source, path, and instrument.)
A seismogram shows s and the station is 120 km from the earthquake. What does this imply about ? Is this consistent with typical crustal rock?
Why does an SH wave not convert to a P-wave when it reflects from a horizontal interface, while an SV wave can? Answer using particle motion geometry.
| Wave | Motion | Speed | Exists in |
|---|---|---|---|
| P | Longitudinal (∥ ray) | Solids + fluids | |
| S | Transverse (⊥ ray) | Solids only | |
| Rayleigh | Retrograde ellipse (P+SV) | Any half-space | |
| Love | Horizontal SH | Layered only |
always. S-waves require (shear restoring force).
Next class: Lab 1 — Introduction to Python — computing , for different rock types
Lecture 6 (Apr 6): Wavefronts, Rays, and Snell's Law
Source: Wikimedia Commons — Public Domain
Instructor note: If the student who experienced the 2001 Nisqually earthquake (M6.8) from Tacoma is present, invite them to describe what they felt — the succession of sharp jolt, strong shaking, and rolling motion maps directly onto P, S, and surface wave arrivals. They indicated willingness in the intake survey.