Earthquakes: What They Are, Why They Happen, and Whether We Can Predict Them

## The three tectonic boundaries that make earthquakes

Understanding why earthquakes happen where they do requires understanding how plates interact. There are three boundary types, and each produces a distinct seismic signature.

Convergent boundaries — plates moving toward each other. These are the most dangerous. When two plates collide, one is usually forced beneath the other in a process called subduction. The descending slab drags against the overriding plate, locking for centuries and then rupturing along hundreds of kilometres at once. Every earthquake above magnitude 9.0 in recorded history has occurred at a subduction zone. Japan sits atop several of them, where the Pacific, Philippine, Eurasian, and North American plates interact — which is why Japan experiences more major earthquakes than almost any other nation on Earth.

The 2004 Indian Ocean earthquake, magnitude 9.1–9.3, was a subduction-zone event: thrust faulting on the seafloor lifted the water column and generated a tsunami that killed close to 300,000 people across ten countries in Southeast Asia and East Africa.

Transform boundaries — plates sliding past each other. Two plates grind horizontally, building immense friction along a near-vertical fault. California's San Andreas Fault is the world's most famous example, and the 1906 San Francisco earthquake it produced destroyed the city and effectively founded modern earthquake engineering. Transform events are usually shallower and rarely exceed magnitude 8.

Divergent boundaries — plates pulling apart. Two plates separate, thinning the crust and creating a rift valley. These generally produce smaller, shallower earthquakes. The East African Rift System — running through Ethiopia, Kenya, Tanzania, Uganda, Rwanda, and down toward Mozambique — is the largest active continental divergent boundary on the planet, and it is the reason East Africa is measurably more seismic than West Africa.

## What actually travels through the ground

An earthquake is not one shake — it is a sequence of wave types arriving at different speeds, and that difference is what makes early warning possible at all.

P-waves (primary waves) travel at roughly 6 kilometres per second. They arrive first, cause a small rapid back-and-forth motion, and do comparatively little damage. Crucially, they are detectable and measurable.

S-waves (secondary waves) travel at roughly 3.5 kilometres per second. They arrive after the P-waves and shake the ground side to side with far greater force.

Surface waves (Love and Rayleigh waves) are slowest and arrive last. They produce the visible rolling motion that flattens buildings and is responsible for most structural collapse.

Early warning systems detect the harmless P-wave, estimate the magnitude within seconds, and broadcast an alert that outruns the destructive S-waves and surface waves — because a radio or internet signal travels at close to the speed of light while seismic energy crawls at a few kilometres per second.

## Can earthquakes be predicted? The honest answer

Researchers have documented several phenomena that sometimes precede major earthquakes, and none of them are reliable enough to evacuate a city on.

Foreshocks. Many great earthquakes are preceded by smaller ones nearby — the 2011 Tohoku event followed a magnitude 7.2 foreshock two days earlier. But the overwhelming majority of small earthquakes are not foreshocks, and the distinction is only visible in hindsight.

Ground deformation. GPS networks and InSAR satellite radar now measure surface movement to centimetre accuracy. Slow-slip events, where plates creep rather than snap, can be detected weeks in advance — but they do not tell you where or when a rupture will start.

Radon gas emissions. Stress fractures can release trapped gas, and some studies detect radon spikes before events. Results are inconsistent and remain contested among seismologists.

Animal behaviour. Anecdotes are abundant; controlled evidence is not. Animals likely feel P-waves before humans do, which is seconds of warning, not days.

The scientific consensus is unambiguous: no method currently exists that can reliably state the time, place, and magnitude of a future major earthquake with enough notice to move populations. What science does well is long-run probabilistic hazard — "a 60% chance of a major rupture on this fault in 30 years" — and short-run warning measured in seconds.

## Where the science genuinely succeeded: warning in seconds

Japan's Seafloor Observation Network for Earthquakes and Tsunamis (S-Net) runs thousands of kilometres of undersea cable and sensors directly over the offshore subduction zone. It is the most sophisticated system on Earth, adding roughly 20 seconds to earthquake warning times and cutting as much as 20 minutes off tsunami warnings.

Twenty seconds sounds trivial. It is enough to brake bullet trains automatically, open fire-station doors before the frames warp shut, shut gas pipelines, halt a surgery mid-incision, stop lifts at the nearest floor, and get a family under a table.

California's MyShake app took the opposite approach: instead of dedicated hardware, it turns smartphone accelerometers into a distributed sensor grid. Since 2019 it has issued 6.8 million alerts across 194 earthquakes.

The February 2023 Kahramanmaraş sequence in Turkey was the sharpest test of the concept. Retrospective analysis showed detection solutions in about 4 seconds for the magnitude 7.8 Pazarcık rupture and about 10 seconds for the magnitude 7.5 Elbistan event — enough for up to 20 seconds of warning in the worst-shaken areas. Turkey had no functioning national warning system. More than 50,000 people died.

## The economics that reach Africa

Earthquakes do not stay local in a commodity-linked world. Japan is one of the largest importers of crude oil, liquefied natural gas, and agricultural commodities on the planet, and a major seismic event that idles refineries, ports, and factories moves global demand within days. After the 2011 Tohoku earthquake, Japanese oil imports fell roughly 15% in the following weeks. Nigeria, whose federal budget assumptions are anchored to a crude benchmark, feels that as a price signal: Brent is currently trading around *** (live)*, and a demand shock in a top-five importer is transmitted straight into the revenue line that funds Nigerian state budgets.

That is the underappreciated African exposure. West Africa's direct seismic risk is low; its indirect exposure — commodity prices, shipping insurance, and Indian Ocean tsunami reach along the East African coast — is not.