How Do Earthquakes Work? Tectonic Plates, Seismic Waves, and Prediction

The Earth's surface is cracked into 15 major tectonic plates that drift between 1 and 10 centimeters per year. When those plates grind past each other, collide, or pull apart, the stored energy releases as an earthquake. About 500,000 detectable earthquakes occur annually, 100,000 can be felt, and roughly 100 cause real damage. Despite decades of research, we still cannot reliably predict when or where the next big one will strike.

What's Actually Happening Underground

Tectonic plates float on the asthenosphere -- a layer of semi-molten rock about 100-200 km below your feet. Convection currents in the Earth's mantle push these plates around like rafts on a slow-motion river. Where plates meet, stress builds along faults (fractures in the rock) until the friction holding them in place fails.

That sudden slip is an earthquake. The energy radiates outward from the rupture point (the hypocenter, or focus) in all directions. The point on the surface directly above is the epicenter -- the spot that shows up on news maps. You can simulate this process and see how seismic energy propagates through different terrain types.

Three Types of Faults

Not all earthquakes are the same. The type of fault determines how the ground moves:

• Normal faults: Plates pull apart. One side drops down relative to the other. Common at mid-ocean ridges and rift zones like East Africa.
• Reverse (thrust) faults: Plates push together. One side rides up over the other. These produce the most powerful earthquakes -- the 2011 Tohoku quake (magnitude 9.1) was a thrust fault event.
• Strike-slip faults: Plates slide horizontally past each other. The San Andreas Fault is the most famous example, with the Pacific Plate grinding northwest past the North American Plate at about 46 mm per year.

Seismic Waves: The Three Shakes

When a fault ruptures, it generates three distinct types of seismic waves, each with different speeds and effects:

P-waves (Primary) arrive first. They're compressional waves -- they push and pull rock in the direction they travel, like sound waves through air. P-waves move at 5-8 km/s through the crust and can travel through solids, liquids, and gases. They feel like a sudden thud or boom.

S-waves (Secondary) arrive next. They shear rock side-to-side or up-and-down, perpendicular to their direction of travel. S-waves are slower (3-5 km/s) but carry more energy. They can't pass through liquids -- which is how we know Earth's outer core is molten. These are the waves that knock things off shelves.

Surface waves arrive last and do the most damage. Love waves shake the ground horizontally, while Rayleigh waves roll the surface in an elliptical motion -- like ocean waves passing through solid ground. Buildings aren't designed for this kind of motion, which is why surface waves cause the most structural failures.

Richter vs. Moment Magnitude: Measuring Earthquakes

The Richter scale, developed in 1935, measured the amplitude of seismic waves on a specific type of seismograph. It worked well for local, moderate earthquakes in Southern California, but it broke down for very large or very distant events.

Since the 1970s, seismologists have used the moment magnitude scale (Mw) instead. It measures the total energy released based on the fault area, the amount of slip, and the rigidity of the rock. Both scales are logarithmic -- each whole number represents roughly 32 times more energy. A magnitude 7 releases about 1,000 times more energy than a magnitude 5.

The largest earthquake ever recorded was the 1960 Valdivia earthquake in Chile at magnitude 9.5. It released energy equivalent to 2.7 billion tons of TNT and generated a tsunami that crossed the entire Pacific Ocean.

Can We Predict Earthquakes?

The short answer: no. Not with any useful precision. Despite billions of dollars in research, no one has demonstrated the ability to predict the specific time, location, and magnitude of an earthquake before it happens.

There have been plenty of claimed precursors -- unusual animal behavior, radon gas emissions, foreshock patterns, groundwater changes. None have proven reliable enough for actionable predictions. The 1975 Haicheng earthquake in China was famously "predicted" based on foreshocks and animal behavior, leading to an evacuation that saved thousands. But the next year, the Tangshan earthquake killed over 240,000 people with no warning at all.

What we can do is forecast probabilities. The USGS estimates a 72% chance of a magnitude 6.7+ earthquake hitting the San Francisco Bay Area before 2043. That's useful for building codes and emergency planning, but it won't tell you to leave town next Tuesday.

Early Warning Systems: Seconds That Save Lives

What we do have are early warning systems, and they're genuinely useful. Japan's system detected the 2011 Tohoku earthquake and sent alerts to Tokyo -- 80 seconds before the shaking arrived. That was enough time to stop bullet trains, open fire station doors, and send millions of phone alerts.

The system works because electronic signals travel faster than seismic waves. Sensors near the epicenter detect P-waves (which arrive first but cause less damage) and calculate the likely magnitude within seconds. Alerts reach distant cities before the destructive S-waves and surface waves arrive. The US ShakeAlert system now covers the West Coast with similar capabilities.

Understanding earthquake mechanics makes earthquake simulations much more meaningful -- you can see how magnitude, depth, and distance interact to determine what you'd actually feel. Earth's geology connects to other natural forces too: volcanic eruptions often occur at the same plate boundaries, and the energy scales involved rival asteroid impacts. Even tornadoes, while driven by atmospheric rather than geologic forces, share that same theme of nature's power operating on scales that dwarf human engineering.

The next major earthquake isn't a matter of if, but when. We can't stop it, and we can't predict it with precision. But we can build smarter, plan better, and understand the physics well enough to know what's coming in those critical seconds after the fault slips. Sometimes, that's enough.