TREMR

At 7:19 on the morning of September 19, 1985, a Thursday, Mexico City was waking up. People were heading to work, children to school. The epicentre of what was about to happen was far away — on the Pacific coast of Michoacán, where the Cocos tectonic plate was diving under the North American plate 400 kilometres to the west. In a typical earthquake, distance is protection. This one would not work that way.

The magnitude 8.1 rupture lasted about two minutes. By the time the seismic waves reached the capital they had been travelling for nearly a minute — and then they hit the ground beneath Mexico City, and everything changed. In parts of the city, the ground shook for more than three minutes. Somewhere between 5,000 and 10,000 people died, possibly more: the official figures were disputed from the beginning, and independent estimates have run considerably higher. More than 400 buildings collapsed outright. Another 3,000 were damaged beyond use. Thirty thousand people lost their homes in a single morning.

The Lake That Wouldn't Let Go

To understand why the earthquake was so destructive in a city 400 kilometres from its epicentre, you need to know what Mexico City is built on. The Aztec city of Tenochtitlán was built on an island in the middle of Lake Texcoco. When the Spanish conquered it in 1521, they began draining the lake — a process that continued for centuries until almost nothing remained. What was left under the modern city was a thick layer of extraordinarily soft, water-saturated clay: the sediment of the old lake bed, up to 40 metres deep in places.

That clay behaves almost like a liquid during an earthquake. When seismic waves enter it, they slow down dramatically — and because energy is conserved, slowing waves grow taller. The amplitude of ground motion in the lake-bed zone was amplified by a factor of somewhere between 10 and 50 compared to the surrounding hills. A city built on solid rock nearby might have swayed for thirty seconds. Parts of Mexico City moved for over three minutes.

While the Government Hesitated

The government's initial response was slow. President Miguel de la Madrid was slow to appear publicly, slow to accept international assistance, and slow to coordinate a rescue effort commensurate with the scale of what had happened. The PRI — the Institutional Revolutionary Party — had governed Mexico without interruption for more than fifty years, and its instinct was to manage information and maintain control rather than open the disaster to outside involvement.

Into that gap stepped the citizens of Mexico City. Within hours, tens of thousands of volunteers had organised themselves into rescue brigades — brigadas de rescate — digging through rubble with their hands, passing debris down human chains, listening for voices. They coordinated without central direction. They organised their own food and supply networks. They worked alongside the small number of international search-and-rescue teams that eventually arrived. Many of the people pulled alive from the rubble in the days after September 19 were found by volunteers, not by the government.

The earthquake is often credited with triggering a transformation in Mexican civic life. The experience of organising collectively to do what the government could not — or would not — do was formative for a generation of Mexicans. Civil society organisations that formed in the aftermath of 1985 became the nucleus of a broader political mobilisation that, over the following decade, eroded the PRI's grip on power. The earthquake did not create Mexican democracy, but many historians argue it planted seeds that the 1988 election controversy and the 1994 Zapatista uprising would eventually bring to flower.

September 19, Again

On September 19, 2017 — exactly 32 years after the 1985 earthquake, to the day — Mexico City experienced another major earthquake. The M7.1 struck at 1:14 in the afternoon, killing 369 people and collapsing dozens of buildings. The date was already marked: every year on September 19, Mexico City runs a national earthquake drill in memory of 1985. In 2017, the drill had finished just hours before the real earthquake struck.

The coincidence was almost too strange to believe. A city running a rehearsal for the event that was already coming. Some Mexicans took it as a sign; seismologists pointed out that September 19 has no special geological significance, that the 2017 earthquake had a completely different source mechanism from 1985, and that with enough years of history any date will eventually accumulate tragedies. That is true. It is also true that both earthquakes hit on September 19, and that Mexico City was better prepared for the second one in part because of the lessons — the hard lessons, the expensive lessons — of the first.

What the Clay Taught Engineers

The 1985 earthquake fundamentally changed how engineers think about site effects in earthquake design. The selective destruction — some blocks of Mexico City devastated while neighbouring blocks were untouched — was a visible demonstration of how much the ground beneath a building matters, sometimes more than the building itself. The concept of soil amplification and resonance, which seismologists had studied theoretically, was now confirmed at catastrophic scale in a modern capital city.

Mexico overhauled its building codes after 1985. It created a network of seismic sensors — the Sistema de Alerta Sísmica Mexicano — that can detect large earthquakes off the Pacific coast and send a warning to Mexico City before the waves arrive: typically 60 to 120 seconds of notice, which is not much but is enough to stop an elevator, get under a desk, or move away from a window. In 2017, that system worked. The earthquake that struck on September 19 came with a warning. Not enough warning to prevent all the deaths, but enough to save some. That, too, is a legacy of 1985.