Earth's Structure, Plate Tectonics, Earthquakes and Volcanoes

The ground beneath your feet is the outside of a layered planet. Earth's radius, from the surface to the center, is about 6,371 km. The layers differ in their materials and how they behave.

Four layers, from outside in

The crust can be 70 km thick or more beneath high mountains. Even there, it is thin compared with the whole planet. The mantle is the thickest layer and takes up about 84% of Earth's volume.

Hot does not always mean liquid

Almost all mantle rock is solid. Over millions of years, it can flow slowly. A glacier offers a useful comparison: solid ice can creep downhill without becoming a river of water. Magma, or molten rock, forms in particular places in the mantle and crust.

The outer core is completely liquid. Movement in its liquid iron produces Earth's magnetic field. That field helps a compass needle line up roughly north–south.

The inner core is hotter still, about 5,000–6,000 °C. This is similar to the Sun's surface temperature. Huge pressure keeps the inner core solid despite the heat. Moving inward through Earth, pressure and density increase, and temperatures become higher.

Why Earth has layers

Young Earth was very hot and partly molten. Dense iron and nickel sank toward the center. Less dense rocky material remained above them. This sorting produced a metal core with rocky layers outside it.

Near the surface, temperature commonly rises about 25–30 °C per kilometer. Suppose the surface rock is 15 °C, with a rise of 25 °C per kilometer. At 3 km deep, the increase is 75 °C, so the temperature is about 90 °C. Deep mines need cooling. This near-surface rate cannot be extended all the way to the core.

Picture the thin crust

For continental crust 35 km thick, divide its thickness by Earth's radius:

That is about half of 1%. A globe 30 cm across has a radius of 15 cm. Its matching crust thickness would be about 0.08 cm, less than 1 mm.

The deepest drilled hole reached only about 12 km in northwest Russia. That is about 0.2% of the distance to Earth's center. Scientists study the deeper layers indirectly. Earthquake waves, meteorites, and mantle rocks brought up by volcanoes provide clues.

A tectonic plate is thicker than the crust. To understand it, we need two ways of describing Earth's outer layers: their materials and their behavior.

Two ways to divide the same rocks

Crust, mantle, and core describe different materials. Lithosphere and asthenosphere describe how the outer rocks behave.

The lithosphere is the cool, rigid shell. It includes all the crust and the uppermost mantle, joined together. It is usually about 50–200 km thick. Its broken pieces are tectonic plates.

The asthenosphere lies beneath the lithosphere, within the upper mantle. It is hotter and softer. Its mostly solid rock can flow slowly, letting the rigid plates move above it. The boundary between these two layers is inside the mantle. It is below the boundary between crust and mantle.

Imagine labeling one rock by its material and its behavior. It can be mantle rock and also belong to the lithosphere. There is no conflict: the labels answer different questions.

Compare the two kinds of crust

Continental crust is thicker but less dense. It sits higher, helping continents stand above sea level. Denser oceanic crust sits lower. A shallow sea can cover continental crust, so water above a place does not always mean oceanic crust below it.

Oceanic crust is continually made and recycled. Continental crust can survive much longer, which explains the very old rocks found on continents.

Compare equal volumes

One cubic meter of granite has a mass of about 2,700 kg. The same volume of basalt has a mass of about 3,000 kg. Basalt has 300 kg more mass in the same space:

So basalt is about 11% denser. Density is mass for a given volume, not simply the size or total mass of an object.

When an old oceanic plate meets a continental plate, the oceanic plate sinks underneath. Its density depends on the whole lithosphere. Thin basalt crust plus cold, dense mantle rock make old oceanic lithosphere denser overall. Comparing crust alone leaves out much of the sinking plate.

Continents have not always been in their present positions. In 1912, Alfred Wegener proposed continental drift: continents slowly move across Earth's surface. His 1915 book brought together several kinds of evidence.

A former joined continent

Wegener called the joined landmass Pangaea, meaning “all lands.” It existed about 300–200 million years ago. Breakup began about 200 million years ago. The Atlantic opened as the Americas moved away from Europe and Africa.

South America's eastern edge and Africa's western edge look like matching puzzle pieces. The fit improves when their submerged continental shelves are matched. Today's coastlines change as waves wear land away and sediment builds it up. The shelf edges better show the continents' outlines.

Fossils across an ocean

Fossils of Mesosaurus, a reptile of lakes and coastal waters, occur in South America and southern Africa. An animal like this could not cross thousands of kilometers of open ocean. Joined land explains how its remains occur on both sides.

The seed fern Glossopteris gives a wider pattern. Its fossils occur in South America, Africa, India, Antarctica, and Australia. Bringing these lands together also brings the fossil sites together.

Rocks, mountains, and old climates

Rocks of the same type and age match across oceans. The Appalachians in North America line up with related mountains in Greenland, Ireland, Scotland, and Norway when the Atlantic is closed on a map. A mountain chain that was later split fits these observations.

Glaciers leave scratches and rocky deposits. Marks about 300 million years old occur in parts of India, Africa, South America, and Australia. Rejoining the continents makes sense of ice spreading across lands once near the South Pole.

Coal records a different climate: it forms from buried swamp plants. Coal in Antarctica points to an earlier, warmer setting. Coal in Britain and the eastern United States formed in tropical swamps, when these lands were nearer the equator.

Put independent clues together

Consider South America and Africa:

1. Their shelf edges fit.
2. Their fossils match across the Atlantic.
3. Matching rock layers connect when the lands are joined.

One explanation accounts for all three: the continents were once together and later separated. Shape alone is weaker evidence because unrelated shapes can seem to fit.

Early explanations sometimes used sunken land bridges to connect fossil sites. These did not account well for the wider rock, mountain, and climate patterns. Light continental rock also cannot simply sink through dense ocean floor. Pangaea was one of several supercontinents in Earth's history; plates are still carrying continents today.

Wegener had strong evidence that continents moved, but his suggested cause failed. Later discoveries showed that ocean floor moves too.

Why the idea was rejected

Wegener suggested continents plowed through ocean floor, driven by Earth's spin and tides. Those forces were too weak. Most geologists rejected drift because it lacked a workable mechanism. Wegener died in Greenland in 1930, before the later evidence appeared.

In the late 1920s, Arthur Holmes suggested slow mantle currents could carry continents. At the time, evidence for this process was limited.

Mapping beneath the oceans

After World War II, ships used sonar to measure ocean depths. Sound traveled down, bounced off the bottom, and returned. The travel time revealed the depth.

Marie Tharp and Bruce Heezen produced detailed ocean-floor maps in the 1950s. Long mountain chains, called mid-ocean ridges, crossed the oceans. Deep, narrow trenches lay elsewhere. The Mariana Trench reaches almost 11 km below sea level.

Ridge shapes differ. Slow-spreading ridges such as the Mid-Atlantic Ridge have a central rift valley. Fast-spreading ridges such as the East Pacific Rise usually have a raised crest. An underwater mountain chain need not have one shape everywhere.

New floor forms and spreads

In the early 1960s, Harry Hess proposed seafloor spreading. Mantle rises at a ridge, and magma forms new oceanic crust. The new floor moves away on both sides. Old oceanic lithosphere sinks back into the mantle at trenches. Formation and recycling explain why spreading does not make Earth keep growing.

Two patterns support spreading:

• Rocks are youngest at the ridge and older farther away on either side.
• Sediment is generally thin near the ridge and thicker on older floor. Older surfaces have had longer to collect mud and shells.

For rock 500 km from a ridge, moving away at 2.5 cm per year, find its age. Convert distance first:

Then divide by the one-side speed:

The estimated age is 20 million years.

From drift to a wider theory

Maps also showed earthquake and volcano belts along ridges, trenches, and mountain ranges. By the late 1960s, these discoveries came together as plate tectonics. Whole pieces of lithosphere move, carrying both continents and ocean floor.

A hypothesis is a testable explanation. A scientific theory is a broad explanation supported by many tests and kinds of evidence. Calling plate tectonics a theory describes the strength and reach of its explanation; it does not mean a guess.

A tectonic plate is a rigid slab of lithosphere. Earth's plates fit together around the whole surface, like pieces of a cracked shell. Their meeting places are plate boundaries.

Plates carry land and ocean floor

Common maps show seven large plates: Pacific, North American, South American, Eurasian, African, Antarctic, and Indo-Australian. Many maps separate the Indian and Australian plates. Smaller plates include Nazca, Cocos, Caribbean, Philippine Sea, Arabian, and Juan de Fuca.

Most plates carry both continental and oceanic crust. The North American Plate includes a continent and much of the western North Atlantic floor. The Pacific Plate is almost entirely oceanic. A continent moves as part of a plate; it does not push through ocean floor by itself.

Coasts and plate boundaries are different lines. The eastern United States lies within the North American Plate. The Mid-Atlantic Ridge is the boundary far offshore. In contrast, South America's western coast lies close to a plate edge. This helps explain why earthquake activity differs between coasts.

Measure movement today

Most plates move about 2–10 cm per year. Some move more slowly. Across the East Pacific Rise, the plates separate at more than 15 cm per year. A separation rate describes the change in distance between plates; a one-side rate describes motion away from the ridge on one side.

GPS stations fixed to the ground measure small position changes over time. A few centimeters in a year seems tiny. Repeating that movement for millions of years carries continents across oceans.

Calculate distance

Use distance = speed × time. At 5 cm per year for one million years:

There are 100,000 cm in 1 km:

In 100 million years at that speed, the distance is 5,000 km. Always keep the unit attached to the answer. Five million centimeters is not five million kilometers.

Calculate time

Now consider an ocean that reaches a width of 5,000 km at a steady widening rate of 2.5 cm per year. First convert the width to 500,000,000 cm. Then use time = distance ÷ speed:

This simple Atlantic example gives about 200 million years, matching the broad timescale of Pangaea's breakup. Plate rates can change through Earth's history, so present rates give estimates over long periods. The Atlantic is still widening, while subduction around the Pacific removes ocean floor and gradually reduces that basin.

Earth's moving surface is linked to its hot interior. Heat and gravity act together through slowly flowing mantle rock and the plates themselves.

Where the heat comes from

Some heat remains from Earth's formation. More is released as radioactive elements in rocks decay, including uranium, thorium, and potassium. Heat slowly escapes toward the surface.

Convection moves heat as material rises and sinks. Hotter material expands and becomes less dense. It rises. Cooler, denser material sinks. Together, these movements can form a convection current.

Soup heated from below shows this movement quickly. The mantle behaves very differently in time and state: it is mostly solid and flows only a few centimeters per year. Over millions of years, its hot rock rises, spreads, cools, and sinks. Plates form the cool outer part of this connected system. They are not passengers on a fast liquid river.

Gravity moves plates from ridges

At a mid-ocean ridge, hot new lithosphere stands high. Farther from the ridge, lithosphere cools, becomes denser, and sits lower. This makes a gentle slope.

Gravity lets the plate slide away from the raised ridge. This force is called ridge push. The name can suggest magma pushing sideways like a piston, but the key movement is gravity acting on the slope.

Trace the cause in order: newly formed lithosphere is hot; hot lithosphere stands high; cooling lowers the older floor; gravity moves the plate away from the high ridge.

A sinking slab pulls

At a trench, old oceanic lithosphere is cold and dense. Its edge sinks into the mantle. This sinking section is the slab. As it descends, it pulls the attached plate behind it. That force is slab pull.

A hanging tablecloth offers a simple comparison. As more cloth hangs over the edge, its weight pulls more cloth off the table. The comparison shows the pull from a sinking part; actual lithosphere is rigid rock moving through a slowly flowing mantle.

Many scientists regard slab pull as the strongest plate-driving force. Plates attached to long sinking slabs tend to move quickly. Their movement also affects mantle flow beneath and around them.

Keep the mechanisms distinct

The core is not directly shoving the plates. Heat from Earth's interior supports this whole system, while differences in density and gravity produce movement. A liquid-water model can show rising and sinking, but it changes state and speed compared with the solid mantle.

At a divergent boundary, plates move apart. It is also called a constructive boundary, because new crust forms. The process happens both under oceans and within continents.

Make ocean floor at a ridge

Hot mantle rock rises beneath separating plates. As it rises, the pressure on it falls. Some rock starts melting without extra heating. This is decompression melting: a pressure drop allows partial melting.

Magma rises and cools into basalt, making new oceanic crust. Below that crust, mantle rock cools and becomes rigid. Together, the crust and rigid mantle make new oceanic lithosphere. A plate is therefore not made entirely from cooled lava.

The raised chain where this happens is a mid-ocean ridge. Most divergent boundaries lie underwater. The Mid-Atlantic Ridge runs through the Atlantic, extending from the Arctic toward Antarctica. Underwater basalt lava often cools into rounded pillow shapes.

New crust forms along the ridge while older floor moves away on both sides. A cross-section shows the youngest rocks in the center. Ages increase outward in a matching pattern. Sediment also has longer to collect on the older floor.

Watch a ridge above sea level

Iceland sits on the Mid-Atlantic Ridge. The North American and Eurasian plates pull apart across the island. A hotspot provides extra magma, helping the ridge stand above the sea.

Long cracks, called fissures, allow eruptions across the ground. Basalt lava is usually runny. Gas escapes fairly easily, so many eruptions are gentle lava flows. Frequent earthquakes are mostly shallow and fairly small, because the lithosphere is thin and hot.

Split a continent

Stretching also breaks continental crust along faults. Some blocks drop between the faults, forming a rift valley with steep sides. This is happening in the East African Rift.

If the rift continues widening, the continent can split. The sea can flood the gap and a new ocean basin can develop. The Red Sea formed as Africa and Arabia separated. A rift valley and a mid-ocean ridge are different stages or settings of plates moving apart.

Read a cross-section

Suppose arrows point away from a central ridge. Basalt at the ridge is young, and older basalt lies on both sides. The arrows show separation; the rock ages show new floor forming between the plates.

An empty hole does not remain between them. Rising magma supplies new crust as separation continues. In a continental rift, the valley is lowered land bounded by faults, not a crack opening all the way to the core.

At a convergent boundary, plates move toward each other. Their crust types help determine whether a plate sinks or the crust thickens into mountains.

Subduction returns a plate to the mantle

Subduction means one plate sinks beneath another. A deep ocean trench marks where the slab bends downward. Old, cold oceanic lithosphere is dense enough to sink. Where lithosphere is removed, the boundary is also called destructive.

The Juan de Fuca Plate also sinks beneath western North America, near British Columbia, Washington, and Oregon. This subduction supplies magma to Cascade volcanoes, including Mount St. Helens.

Why volcanoes form above a slab

Water is carried down inside the sinking plate's rocks. At roughly 100 km deep, water is released into hot mantle above the slab. It lowers the mantle rock's melting point. Some of that rock melts.

The magma rises through the upper plate, feeding volcanoes away from the trench. The main magma source is the mantle above the slab. Simply picturing the whole sinking plate melting misses the role of water.

At oceanic–oceanic boundaries, this builds a curved island arc. The Aleutian Islands and Mariana Islands are examples. The Mariana Trench is almost 11 km deep. At oceanic–continental boundaries, volcanoes form on land, as in the Andes.

Two continents collide

Continental lithosphere is too buoyant to sink far. When continents meet, their rocks fold, break, and thicken. Thickened crust rises into mountains. India began colliding with Asia about 50 million years ago, building the Himalayas.

This collision continues, with parts of the range rising more than 1 cm per year. Collision zones have strong earthquakes but very few volcanoes. They lack the sinking oceanic slab that supplies water for the usual subduction-related melting.

Follow earthquakes and eruptions

Subduction produces shallow earthquakes near the trench and deeper earthquakes farther beneath the upper plate. Some reach about 700 km deep. The largest recorded earthquake, magnitude 9.5 in Chile in 1960, occurred at a subduction zone.

Thick, sticky magma at subduction zones often traps gas, producing explosive eruptions at steep composite volcanoes such as Mount Fuji.

Some plate edges move sideways past each other. These are transform boundaries, also called conservative boundaries. They shift existing lithosphere without making or removing it.

Sliding can stop, then jump

A transform boundary follows a large fault: a break along which rock moves. The plates' overall motion continues, but friction can lock a section of fault. Nearby rock bends and stores energy.

Eventually the stress is enough for sudden slip. Energy spreads as earthquake waves. This explains why plates moving a few centimeters per year can cause a rapid, powerful earthquake.

Earthquakes here are mostly shallow. Some are large and damaging. A transform boundary does not have the rising mantle of a spreading ridge or the sinking slab of a subduction zone. Sideways sliding itself does not supply magma, so volcanoes are uncommon.

Examples on land and underwater

The San Andreas Fault in California separates the Pacific and North American plates. The Pacific Plate moves northwest relative to North America at about 5 cm per year overall. Nearby faults share some of the movement; a single fault does not carry the entire rate everywhere.

The Alpine Fault on New Zealand's South Island and Turkey's North Anatolian Fault are other examples of major sideways motion. Real boundaries can have extra squeezing or stretching as well as sliding.

Most transform faults are underwater. They connect offset sections of mid-ocean ridge. On a map, the ridge can look like a zigzag, with short connecting faults crossing between ridge sections. A transform segment does not create new floor just because it lies next to a spreading ridge.

Estimate movement stored at a locked fault

Suppose plates move past each other at 4 cm per year. A section stays locked for 150 years. The accumulated movement is:

Convert to meters:

This simple estimate shows why a single slip can move the ground several meters. A fence, road, or stream crossing the fault can become offset sideways. It does not give the date of the next earthquake or guarantee one slip will release all the strain.

Recognize the setting

Look for movement arrows running along the boundary. They show relative sideways motion. Long fault valleys and offset streams support that reading.

A trench points to subduction, while new basalt at a ridge points to divergence. Transform boundaries are identified by sideways motion and faults, not by volcanic cones. Earthquakes alone cannot distinguish the three boundary types, because all can have them.

Earthquakes and volcanoes form patterns across the world. A plate map helps connect those patterns with motion beneath them.

Start with the map key

Boundary lines mark plate edges. Movement arrows show how plates move relative to each other:

• Arrows away from the boundary: divergence.
• Arrows toward the boundary: convergence.
• Arrows along the boundary: sideways, or transform, motion.

Many maps draw triangles along subduction zones. The triangles point toward the plate above the sinking slab. Check the map key before reading symbols; a line without its key gives less information.

Coastlines outline land, not plates. Many boundaries cross oceans, and many coasts lie inside plates. Use the boundary symbols and motion arrows, not the shape of a shore.

Find the main earthquake belts

Most earthquake locations form narrow belts that outline plates. Volcanoes follow many of the same belts. This overlap is evidence that plate interactions cause much of the activity.

The Pacific Ring of Fire is a horseshoe-shaped belt around the Pacific's edges. It runs through the Andes, Central America, the Cascades, Alaska, Japan, the Philippines, Indonesia, and New Zealand. Most of it follows subduction zones. About 90% of the world's earthquakes occur in this belt.

The Alpine–Himalayan belt crosses the Mediterranean, Turkey, Iran, and the Himalayas. It has many strong earthquakes. Mid-ocean ridges form another network, marked by mostly shallow earthquakes through ocean basins.

Add depth and landforms

A map of earthquake dots shows where earthquakes occur. A cross-section adds depth. Near a subduction trench, earthquakes are shallow. Farther beneath the upper plate, they become deeper. Their sloping pattern traces the descending slab.

Compare three settings. Japan has offshore trenches, subduction, deep earthquakes, and explosive volcanoes. Iceland has a spreading ridge, shallow earthquakes, and much basalt volcanism. The Himalayas have continental collision, high folded mountains, and few volcanoes.

Earthquakes with no nearby volcanoes can fit either transform motion or collision. Use arrows and landforms to decide between them. A long sideways fault differs from a wide belt of squeezed mountains.

Account for points away from boundaries

Not every dot follows a plate edge. Hawaii's volcanoes lie over a hotspot within the Pacific Plate. Old faults within plates can also move, causing less frequent earthquakes. The New Madrid earthquakes of 1811–1812 occurred in the central United States.

For an unfamiliar map, combine the evidence in order: read the key, check arrows, look for ridges or trenches, and compare earthquake depths. One isolated volcano does not locate a boundary by itself; its setting matters.

An earthquake is sudden ground shaking caused by rock breaking or slipping along a fault. Most occur near plate boundaries, where moving plates stress the rocks.

Store energy, then release it

Friction can lock a fault while plates keep moving. Rock bends and stores energy, like a bent ruler. When the fault slips, that energy spreads as seismic waves. Smaller earthquakes called aftershocks often follow as surrounding rocks adjust.

The focus, or hypocenter, is the underground point where breaking starts. The epicenter is directly above it on the surface. A focus less than about 70 km deep is shallow. Earthquakes can start as deep as about 700 km within sinking slabs.

A shallow earthquake usually produces stronger surface shaking than a deep earthquake of the same size. Distance, local ground, and buildings also affect what happens at a place.

Separate size from local effects

Scientists commonly use moment magnitude, especially for large earthquakes. The older Richter scale was more limited. Modified Mercalli intensity runs from I to XII, from not felt to extreme effects. Intensity often decreases with distance, but ground conditions can change that pattern.

A seismometer detects ground movement. A seismogram is its record. Several stations help locate an earthquake and estimate its depth and magnitude.

Compare energy correctly

Each increase of one magnitude releases about 32 times more energy. An increase of two gives about:

Thus magnitude 7 releases about 1,000 times the energy of magnitude 5. It is not simply two units stronger on an evenly spaced energy scale. A one-unit increase also corresponds to about tenfold recorded wave amplitude under comparable conditions.

The largest recorded earthquake was magnitude 9.5 in Chile in 1960. Magnitude alone cannot tell the damage in every town: depth, distance, ground, and building strength matter.

How a tsunami starts

A large undersea earthquake can suddenly lift or lower the seafloor. This moves the water above it and starts a series of long waves, a tsunami. Landslides and volcanic eruptions can also start tsunamis.

In deep ocean, tsunami waves may be less than 1 m high while traveling about 700–800 km per hour. Near shore they slow and can grow much taller. At 700 km per hour, traveling 1,400 km takes 2 hours. An undersea earthquake does not always cause a tsunami: it must move enough water.

A volcano is an opening where molten rock, gas, and ash reach the surface. It can also mean the landform built by eruptions. Molten rock underground is magma; at the surface it is lava.

Where magma forms

Solid rock partly melts in particular settings. At divergent boundaries, mantle rises and melts as pressure drops. At subduction zones, water from a sinking plate lowers the melting point of mantle above it.

A hotspot is a long-lived source of volcanism beneath a plate, often away from a boundary. Many hotspots are explained by a mantle plume, a rising column of unusually hot rock. Scientists continue to study how these sources form. Yellowstone is a hotspot beneath a continent; Hawaii is an oceanic example.

Magma is less dense than surrounding rock and rises through cracks. Some collects below a volcano in a magma chamber. Dissolved gases form bubbles as rising magma experiences lower pressure. Opening a soda bottle gives a familiar comparison: falling pressure allows bubbles to grow. Expanding gas helps drive an eruption.

Why eruptions differ

Viscosity means resistance to flow. Basalt magma is usually runny. Repeated flows build wide, gently sloping shield volcanoes such as Mauna Loa. Magma richer in silica is often stickier. Gas pressure can build until ash, fragments, and hot gas burst out.

Composite volcanoes, also called stratovolcanoes, build from layers of lava and ash. Mount St. Helens, Mount Fuji, and Mount Pinatubo are examples. Boundary setting helps explain common styles, but it does not make every eruption identical.

Read a hotspot chain

A relatively fixed hotspot can form new volcanoes as a plate moves over it. Volcanoes carried away become older along the plate's direction of movement. In Hawaii, active volcanoes lie at the southeastern end, and islands become older toward the northwest.

Kauai lies about 500 km from the active hotspot and has lavas about 5 million years old. Convert 500 km to 50,000,000 cm. Divide by 5,000,000 years to estimate 10 cm per year toward the northwest, close to the regional plate speed of about 7–10 cm per year.

Describe activity carefully

An active volcano is erupting or likely to erupt again after recent activity. A dormant volcano is quiet but can erupt again. An extinct volcano is not expected to erupt again. These labels have uncertain boundaries: a long quiet period does not establish that a volcano is extinct. Explosive eruptions can produce fast, hot ash-and-gas clouds; flowing lava is only one part of volcanism.

Mountains grow as plates squeeze, stretch, or add volcanic rock to the crust. A whole mountain range usually takes millions of years to build. Erosion begins wearing it down while it rises.

Fold and thicken the crust

At convergent boundaries, compression squeezes rock layers. Hot rock deep underground can bend slowly into folds. Upward arches are anticlines; downward troughs are synclines. Folding and crustal thickening help raise fold mountains.

India's collision with Asia built the Himalayas. Africa and smaller plates pushing into Europe helped build the Alps. The Andes and Rockies also contain much folded rock.

Fossils near Everest's summit record an older setting. These sea creatures lived before their remains became sedimentary rock on an ancient ocean floor. Later collision lifted that rock high above the sea. The fossils do not mean the ocean once stood at the mountain's present height. Everest is about 8,849 m above sea level.

Break and shift rock blocks

Rocks can also break along faults. The direction of stress helps explain their movement.

Stretching breaks crust into large blocks. Some drop into valleys; others stand higher or tilt. These are fault-block mountains, found in the Sierra Nevada and Basin and Range region of the western United States. A normal fault belongs to stretching even though one side of the landscape ends up high.

Build mountains from eruptions

Volcanic mountains grow as lava and ash pile up. The Cascades and many Andes volcanoes form above subduction zones. Kilimanjaro lies beside the East African Rift. Hawaiian volcanoes grow over a hotspot.

Height depends on the starting point of measurement. Mauna Kea rises more than 10,000 m from its base on the seafloor, but about 4,200 m above sea level. A base-to-summit height and a height above sea level answer different questions.

Balance uplift and erosion

Uplift raises the land. Rivers, glaciers, and landslides remove rock. At 1 cm of uplift per year, one million years would produce 1,000,000 cm, or 10 km, of rise if no rock were removed. Mountains do not simply gain that much height because erosion happens too.

The Himalayas began forming about 50 million years ago and remain high and jagged. The Appalachians formed more than 250 million years ago and have been worn into lower, rounder ranges. A large earthquake can raise ground a few meters, but building a range requires many movements and long-lasting uplift.

Magma rising inside a volcano changes the surrounding rock. Scientists monitor these changes to judge whether an eruption is becoming more likely.

Listen for moving magma

Magma can break rock as it moves. Seismometers record the resulting small earthquakes. A sudden increase in their number is an earthquake swarm. Continuous shaking, called tremor, can also reveal underground movement.

A swarm is a sign of unrest. It is useful evidence, but magma can move without reaching the surface. Scientists compare the shaking with other measurements and the volcano's normal behavior.

Measure a changing shape

Rising magma can push the ground outward and upward. This change in shape is ground deformation.

Tiltmeters record tiny changes in slope. GPS stations record how ground positions shift. Radar satellites map changes in height over large areas. A swelling flank can reveal a pressure increase below it, even when nothing is visibly erupting.

Track gas and heat

Fresh magma can release more sulfur dioxide or carbon dioxide. Gas instruments measure the amounts and changes over time. Thermal cameras and satellites detect warmer ground, water, or crater areas. Melting snow and ice can be another clue.

Imagine a volcano with more small earthquakes, rising gas output, and a swelling surface. The measurements agree with magma moving upward. This is a stronger reason to raise concern than a single unfamiliar gas reading by itself.

Make a forecast

A forecast estimates how likely an eruption is. It rarely fixes the exact day, size, or style. Some volcanoes show unrest for months or years without erupting. Monitoring measures a changing natural process; it cannot stop that process.

Before Mount Pinatubo erupted in the Philippines on June 15, 1991, scientists observed earthquakes, gas releases, and small explosions over about ten weeks. Their warnings supported evacuations of tens of thousands of people. The forecasts are estimated to have saved at least 5,000 lives.

Understand the limits

At Mount St. Helens, earthquakes began in March 1980. The north flank bulged by more than 1 m per day. Scientists warned about the danger, but could not name the exact eruption time or sequence.

On May 18, an earthquake triggered a major landslide. Removing the bulging flank released a sideways blast; 57 people died. Warning signs had shown growing danger without revealing every detail of the eruption.

A quiet period does not establish safety. Monitoring is useful because it can give people and authorities time to respond to changes, even when an exact prediction remains out of reach.

Scientists can identify earthquake-prone places and quickly detect earthquakes. Reliably naming a future major earthquake's time, place, and size is still beyond present methods.

Prediction asks for specific details

An earthquake prediction would say when and where an earthquake will happen and how large it will be. Faults do not usually give reliable signs before slipping. Animals, unusual weather, and small tremors have not provided a reliable method for predicting major earthquakes.

Scientists instead make forecasts. They use past earthquakes, ancient fault traces in rocks and soil, plate speeds, and measurements of strain. These support estimates of the chance of an earthquake of a certain size within a region and time period.

Read the whole probability statement

A USGS forecast published in 2016 gave a 72% chance of at least one magnitude 6.7 or larger earthquake in the San Francisco Bay Area during 2014–2043. This historical forecast names a region, size threshold, and 30-year period.

It describes a large earthquake as more likely than not over that whole period. It gives no exact date, year, or fault. The 72% figure is not the chance for each individual year. Earthquakes do not arrive at perfectly regular intervals, so “due tomorrow” is not a valid reading of a long quiet period.

Hazard maps show where strong shaking is more likely. High-hazard areas include many plate boundaries. In the United States, the West Coast, Alaska, and Hawaii stand out, while places such as the central Mississippi Valley still have real earthquake hazard.

Early warning starts after the earthquake

Systems such as ShakeAlert and Japan's nationwide warning system use nearby seismometers to detect an earthquake's first seconds. Electronic messages travel much faster than earthquake shaking. Alerts can therefore reach more distant places before strong shaking does.

The system has detected an event already in progress. It has not predicted a future one. Nearby places may get no useful warning because shaking arrives too soon.

Calculate available warning time

For a simple estimate, take strong shaking traveling at 3.5 km per second. A city 70 km away receives shaking after:

If detection and sending the alert take 5 seconds, the useful warning is:

On receiving a shaking alert, drop, cover, and hold on. Automatic systems can also slow trains. Forecasts help with longer-term preparation, while early warning uses the short gap between detection and arriving shaking.