The shaking started just before dawn, right beneath the ancient Iranian city of Bam. Within moments, the magnitude 6.6 earthquake had torn the city apart. Bam was an archaeological treasure in a sea of empty desert and yet the quake seemed to strike it like a bullseye, says Cambridge geophysicist Professor James Jackson. A third of the city’s population died that day in 2003 – some 34,000 people – and most of its buildings crumbled to rubble, including the famous 2000-year-old citadel.

Jackson had first come to Iran as a PhD student in the late 1970s to help lay early seismometers detecting quakes. When he returned to investigate the Bam disaster, he remembered that another Iranian oasis in the desert, Tabas, had been levelled by a quake not long after his drive across the desert. And there was another town, Sefidabeh, sandwiched between two deserts (one named the desert of death, the other the desert of hell) wiped out just four years before that. So what was going on?

The answer was in the earth – these settlements were all sitting on a “fault” in the rock. Faults let water pool, making a city in the desert possible, but they are also where the ground moves when something else rumbles out of its depths: an earthquake. Cities across the world are built on faults. Some nations, like Japan, are constantly thrumming with quakes.

In Vanuatu, which straddles the infamous Ring of Fire quake hotspot of the Pacific, a 7.3-magnitude earthquake struck in the capital Port Villa on Tuesday. At least 14 people are dead and hundreds injured, including some trapped in the ruins of collapsed buildings and embassies.

Other faults lie quiet for centuries, past disasters fading from memory until the next big one hits. “It’s like a time-bomb,” says Jackson. The earthquakes that shook Turkey and Syria in 2023 were even bigger than the Bam quake, killing almost 60,000 people, and also on a fault.

In Australia, quakes are usually less powerful but surprisingly frequent. There will be one rumbling somewhere most days and a magnitude 5 or more will hit every year or two – though they’re often so remote we don’t feel them. Sometimes, quakes will rattle our cities when they strike right below, as a noisy Melbourne quake did in 2023.

Scientists today usually know where quakes will strike and how hard. They can help engineers build in ways that will save lives, a la the swaying towers of Japan. “What we can’t predict is when,” says Jackson.

So what are earthquakes exactly? Where are the next “big ones” brewing? Why are they so hard to predict? And how do we live with them?

What causes earthquakes?

Earthquakes are, technically, any strong shaking of the Earth. Rocks underground break and slide past each other along faults. “Think of faults like knife cuts in the Earth,” says Jackson. “As rocks move along them, they vibrate.” Pull a brick on a string along a table and it’ll bump and grind along. “It’s those vibrations that cause the trouble.”

When our planet is rocked by big earthquakes, it rings like a bell as shockwaves bounce around the interior (pictured above). Just as X-rays and soundwaves can reveal the shape of whatever they pass through, these “seismic waves” give us a window into the Earth. The denser the material, the faster seismic waves will travel through it. That’s how scientists know what the Earth is made of.

We live on the crust – a rocky skin broken into slow-drifting tectonic plates that fit together like puzzle pieces. These plates make and remake the continents as they jostle. Beneath the crust, the second layer of the Earth – a green, glittery and gooey rock called the mantle – is on the move itself, in a crawl over millions of years as the plates shift overhead, pulling up new crust to the surface and dragging down the old as part of the planet’s vital carbon cycle. Jackson compares it to a conveyor belt.

As the edges of the tectonic plates collide, “what’s at the surface gets scraped off, like a person going up an escalator at the London Underground”. The escalator goes back down underground but the person has to step off or get scraped off.

‘It just compresses and compresses until it can’t any more – it breaks – and the Earth jumps along the fault, say five metres.’

Take the Himalayas, where the edges of two continents are smashing together. “Nepal is being pushed out on top of India in a great ramp,” says Jackson. “Faults are normally held together by friction, so it’s stuck right now. But the whole place is compressing like a rubber ball, and we can measure that. In the case of Nepal, it’s about 15 millimetres a year. It just compresses and compresses until it can’t any more – it breaks – and the Earth jumps along the fault, say five metres.” That’s an earthquake.

Quakes can be contagious. Most are followed by aftershocks (quakes of smaller magnitude) as the fault readjusts following the main shock, usually immediately after. But a big earthquake can trigger another one much further away, even on opposite sides of the globe. The two devastating quakes that hit Turkey and Syria in 2023 were registered as separate events: the first, a magnitude 7.8, struck 95 kilometres from the second, a magnitude 7.7.

Earthquakes can set off deadly tsunamis that roll into shore as the seafloor is warped and thrust up by the quake. If there’s a big enough move in the crust, they can trigger volcanic eruptions too. Likewise, volcanoes can unleash quakes as rising magma fractures rocks and puts stress on surrounding faults.

There are even man-made earthquakes. Scientists say a huge fracking boom across the United States has left an earthquake epidemic in its wake, lifting Oklahoma above California as America’s quake capital. (It’s not the fracturing of shale rock to blame but the injection of wastewater deep underground.) The UK has banned fracking after earthquakes.

Where are the big faults to watch?

If you drive past Lake George down the Federal Highway from Canberra to Sydney, you’re driving over one of Australia’s most active faults. The 75-kilometre line is thought to “slip” 100 metres every million years or so – enough to produce a large quake. But it hasn’t happened in recorded history, and, according to Geoscience Australia, may yet stay dormant for hundreds of thousands of years more. Australia sits on the middle of a tectonic plate, after all, and most of the action happens at the edges.

The world’s biggest hotspot is the Ring of Fire in the nearby Pacific Rim, a 45,000-kilometre horseshoe that runs through places such as Vanuatu, the Philippines, New Zealand as well as the coasts of America and even Russia. Here, many tectonic plates are colliding or slipping under one another.(Beneath Japan alone, three plates are converging.)

Then there’s the big earthquake belt of Eurasia running from the Mediterranean through the Middle East to China where the continent is being “crumpled up like brown paper” as plates collide, Jackson says. “That’s why you have so many mountains.”

Some faults themselves are so big you can see them from space, such as the massive San Andreas Fault through California and the North Anatolian Fault in Turkey. New Zealand is one of the few places on Earth where you can touch a fault – the Alpine Fault, on the South Island, where scientists expect a big quake will hit in the next 50 years. (That fault erupts roughly every 300 years. The last one was in 1717.) And in Iceland, major plates meet above sea level; visitors can stand in a rift valley and gaze up at the cliff edges of North America and Eurasia.

On the surface, faults look like jagged lines but they extend down for kilometres. Mark Quigley, associate professor of earthquake science at the University of Melbourne, thinks of them as planes, complex zones interacting deep underground. Earthquakes might rupture a small part of a fault first, he says, expanding along it at between 2 and 3 kilometres per second. “Sometimes the rupture of an entire fault may take several minutes.”

In Australia, stress does build up into smaller quakes as our tectonic plate inches north-east by about 7 centimetres each year (faster than fingernails grow).

The point where the quake starts is called the hypocentre or “the focus” while the spot directly above it at the surface is the famous epicentre.

The fault that ruptured in Sumatra in 2004, causing a 9.1 magnitude quake and the deadly Boxing Day tsunami, was 1200 kilometres long. “That’s from London to Rome,” says Jackson. And it moved 25 metres, the length of a swimming pool. “That’s huge.”

By comparison, one of the most destructive quakes of Australia’s past – in Newcastle in 1989 – was much smaller, a magnitude 5.6. “And that fault probably only moved centimetres, less than a metre,” says Jackson. “But it was right under the city.”

The same tremor 150 kilometres deep will lose its bite, Quigley says, its energy soaked up by the rocks above before it shakes the ground. “The further you are away, the less chance of you feeling it.” Much depends, too, on the kind of rock, topography and soil in the epicentre area.

On the continents, all quakes are by definition shallow, says Jackson, tens of kilometres down. Meanwhile, though the faults are old, megacities are a new phenomenon. Some of today’s metropolises were tiny towns when the last big quake there struck. “No one remembers any more,” he says. “They have everyday problems – like pollution, poverty, water supply – that take precedence. Then the big quake happens, and it just overwhelms them.”

In Australia, while there are fewer nerves over “a big one” hitting, stress beneath our feet does build up into smaller quakes as our tectonic plate inches north-east by about 7 centimetres each year (faster than fingernails grow). Meanwhile, Jackson and his colleagues theorise that a number of mysterious – and unusually shallow – quakes in our continent’s very centre, could indicate a change in gravity itself. Because of how Australia’s original mountains were ground down over millennia, the rocks beneath are of different density, and so have slightly different gravitational pulls to one another, which appears to build stress.

How do we measure quakes?

There’s the seismometer, that “fancy microphone in the Earth”, as Jackson describes it, listening for the vibrations of quakes. And then there are more exotic instruments.

Jackson was among early earthquake scientists to use emerging military technology – from radar and satellite imagery to GPS (then encrypted and mostly used to guide cruise missiles) – to measure how the Earth had moved after quakes. Indeed, when he started out as a geophysics student in the 1980s, seismology was booming thanks to defence funding. In the Cold War, the only way to monitor nuclear weapons tests by the Soviets was through seismology, for example. A lab where Jackson worked at MIT was funded by the US Air Force, offering access to huge swathes of data from seismometers all over the world. “And every six months or so we had to write a report on [Soviet] nuclear weapons tests in Kazakhstan but none of us were really interested in that. We were there for the earthquakes.”

Likewise, military efforts to map the oceans for submarines helped scientists discover plate tectonics to begin with (solving the mystery of why seafloors were spreading).

‘People want to know the quake will be Tuesday at 3pm. But if you let people think we can predict quakes like that, no one will do anything to deal with them.’

In the 1930s, two geologists in California developed a method for comparing the size, or magnitude, of earthquakes. They were Charles Richter and Beno Gutenberg, though the method is now known only as the Richter scale. Once, this meant actual wiggling needles recording the Earth’s shaking. Today it’s all digital. Each number on the scale represents a 10-fold increase in a quake’s power. A magnitude 6 is 10 times bigger than a 5, for example. And, while most people think of the Richter scale as being a 1 to 10 measure, Richter himself said there was technically no upper limit, only that at the time the highest magnitudes recorded didn’t quite reach 9. But he added, “that is a limitation in the Earth, not in the scale”.

Today the Richter scale has been refined to the “moment” scale, which captures more kinds of seismic waves, not just a quake’s peak vibrations. Still, when one hits, many measurements from sensors all over the world are pulled together, often automatically, to calculate its power. Even “felt scores” from locals on the ground are taken into account.

All this data doesn’t always tell the same story. Government agencies can arrive at different numbers, or take longer than others to calculate. The most powerful quake ever recorded depends on whom you ask – it may have been a magnitude 9.2 in Alaska in 1964 or one just four years earlier off the coast of Chile that was initially logged as magnitude 8.6 but later revised up to a whopping 9.5.

Why are quakes so hard to predict?

In 2009, a magnitude 6.3 earthquake shook central Italy, killing more than 300 people in the medieval town of L’Aquila. Three years later, six Italian scientists were sensationally convicted of manslaughter, accused of failing to warn locals of the risk. They were later exonerated as the science crystallised.

“People want to know the quake will be Tuesday at 3pm,” sighs Jackson. “But if you let people think we can predict quakes like that, no one will do anything to deal with them. And people will die.”

Still, for a long time scientists did try. In the ’70s, after the United States put a man on the moon, “the next big thing they said they’d crack was earthquake prediction”, Jackson says, recalling the dollars poured into research. In Parkland, California, an earthquake had stuck along the San Andreas fault once every 22 years or so, almost like clockwork. But when scientists rigged it with seismometers to catch signs of the next one, it was late by about a decade – and there were no warnings signs.

In 1975, China did famously predict an earthquake. The Chinese Seismological Bureau had been mapping quakes in an earthquake-prone region of the north and noticed one obvious gap. “So they instrumented it to hell,” says Jackson, and detected a swarm of little quakes. “Then suddenly the tremors stopped and that spooked them. They thought that’d be the site of the next big one.” An entire city there, Haicheng, was evacuated to a hillside. “They showed people outdoor movies. And would you believe, almost straightaway a big earthquake happened in the gap. So they saved a lot of lives.” (A few hundred still died – from hypothermia and then a fire while camping in makeshift tents during the evacuation.)

While the ruling Communist Party made a great fuss over their earthquake “prediction”, Jackson says Chinese scientists were warning it was only a lucky guess. “It was pure instinct. Not science.” The next year a second earthquake happened nearby without warning. Some 600,000 people died. More devastating earthquakes would follow. “And finally the Chinese [officials] agreed, OK, this earthquake prediction policy kills people because it’s not true.”

Over the years, scientists have analysed everything from electromagnetic signals rising from the Earth’s interior to animal behaviour to try to find warning signs of earthquakes. None have stood up as hard evidence (although scores of snakes slithering from their hibernation mid-winter out of the quake zone in 1975 was another “sign” that persuaded Chinese authorities to evacuate Haicheng).

‘You still have a quarter million people who are now refugees in their own country … Wouldn’t it be better to just build buildings that won’t fall down?’

Some experts describe earthquakes as a kind of conversation in the rock. As faults move, they change the distribution of stress – certain parts may come under more pressure, others under less. Scientists have tried to calculate the odds of an earthquake striking by measuring how faults have shifted and where pressure is building. But you need great historical data to do that. It’s mostly been useful in predicting aftershocks.

Even if science does find a crystal ball for quakes, Jackson notes that evacuations are not enough. “Say, you predicted the Turkish-Syrian earthquake in February [2023]. You still have a quarter million people who are now refugees in their own country. You have [economic] damage. Wouldn’t it be better to just build buildings that won’t fall down?”

In places such as Chile, New Zealand, Japan and California, quakes hit all the time but don’t cause the same kind of carnage, he says, because people are prepared.

How do we live with earthquakes?

In 2015, Jackson and a host of scientists landed in Kathmandu to talk earthquake preparation. They toured schools newly reinforced against quakes by an entrepreneur (with the support of the Nepalese government, financing from the Australian government, and local labour). “It was very moving to be shown around by the villagers saying, ‘This is my school, I built this one.’ Of course, no one knew then what was coming.”

Two weeks later, one of the worst quakes in Nepal’s history struck that very spot. Nearly 9000 died and 800 schools collapsed, but all 323 of those which had been reinforced stayed standing, Jackson says. In fact, those quake-proof schools would become key co-ordination points during the disaster.

It might not be possible to forecast a quake but it is possible to build to survive it. “We scientists can now say to the architects and engineers, ‘Here’s what you’re up against, this kind of quake, and where, this shaking, this [likely] duration’.”

In places such as New Zealand and California, houses outside the city are often made from wood, light and strong so they won’t collapse. “If I live in a wooden California house outside Los Angeles, I know it’ll be OK after a quake,” he says. “Maybe my hot tub won’t work – that’s my biggest problem.”

Even for all the devastation of New Zealand’s 2011 Christchurch quake – a 6.2 tremor warping roads in a city not known for quakes – most of the 185 people who died were in one building that collapsed. “That tells you how well, in general, they build there,” says Jackson.

In Tokyo, towers dance on shock absorbers, usually thick blocks of rubber, like palm trees blowing in the wind. When Japan was rocked by the most powerful quake in its history in 2011, a magnitude 9, most of the 18,000 people killed died in the resulting tsunami, not the quake itself.

Once a big quake hits near the coast, you have about 20 minutes before a tsunami rolls in. “Every Japanese schoolchild knows you run to high ground,” says Jackson. “And that’s what everyone did in 2011.” On Japan’s Sendai plain, “where there’s nowhere to go”, they climbed specially designed buildings. But the tsunami that came that day was “freakishly” high. “It was the biggest tsunami of the last thousand years. So they weren’t tall enough.”

Neither was the wall protecting the back-up power supply at the Fukushima power plant, resulting in the most catastrophic nuclear meltdown since Chernobyl.

‘Of course, LA and San Francisco are right on the Santa Andreas fault – they get zero seconds warning.’

In some quake-prone countries, intricate early warning systems rigged up to sensors automatically kick in the moment the tremors start. Trains stop and elevator doors open. Water, gas and electricity shut off. “Coastal Japan is about 200 kilometres away from where the quakes start offshore,” explains Jackson. “So you get about 20 seconds warning. In 2011, there were 33 bullet trains going hundreds of [kilometres] an hour across the Sendai plain and every single one was brought to a halt. Of course, LA and San Francisco are right on the Santa Andreas fault – they get zero seconds warning.”

While some institutions take out “catastrophe” bonds to insure nations such as Japan against the next “big one”, experts say a key concern is getting people to follow earthquake building codes, especially in countries where poverty or corruption are rampant. Still, there are success stories.

In Iran, another strong quake hit a city roughly the size of Bam in 2017. This time, the mortality rate was less than 1 per cent (compared to more than 30 per cent during the Bam disaster). New buildings had been following the guidelines.

“You could see these apartments 10 storeys high looking like skeletons, the walls had fallen out but the frames were fine,” says Jackson. “And the people inside survived.”

- with Lia Timson

Our new anthology, Why Do People Queue for Brunch? The Explainer Guide To Modern Mysteries, featuring 26 popular Explainers, is in bookstores now – full of astonishing facts and occasional weirdness for curious minds this Christmas.