Twenty years ago, the seafloor west of Indonesia abruptly pushed upward as a deep undersea fault, where two of Earth’s tectonic plates meet, slipped. The upward shove violently shifted the seawater above, transferring deadly energy from ground to water, and sending the water speeding toward land.

The series of tsunamis generated by that magnitude 9.2 earthquake reached towering heights; Indonesia’s Banda Aceh, close to the epicenter, was engulfed by a wall of water 51 meters tall. The waves killed an estimated 230,000 people across 15 countries, with tens of thousands more reported missing.

It was the deadliest tsunami in recorded history, and one that changed tsunami science.

The Indian Ocean disaster “was a wake-up call,” says Bernardo Aliaga, the head of UNESCO’s tsunami resilience program. Researchers have worked not only to better understand tsunamis, but also to set up warning systems in ocean basins around the world — and to better prepare coastal communities to respond quickly when the alarm sounds.

Disastrous tsunamis have driven change

Tsunamis are generated by movements of the ocean floor that displace and mobilize large amounts of seawater, towers of watery energy that speed toward coastlines. Generally, such ground shifts occur due to the sudden slippage of faults in Earth’s crust. But underwater landslides or massive volcanic eruptions — such as the 2022 eruption of the Hunga-Tonga-Hunga-Ha’apai volcano in the South Pacific — can also generate the towering walls of water (SN: 1/21/22; SN: 8/29/22).

Historically, big advances in tsunami warnings have tended to come only in the wake of devastation.

The Pacific Tsunami Warning Center, based in Hawaii, was the first global tsunami alert system in the world. The United Nations established it in 1965, five years after a magnitude 9.5 earthquake — still the largest ever recorded instrumentally — struck off the coast of Chile. That quake triggered a massive wall of water that swept west across the Pacific. It deluged Hawaii 15 hours later before rushing on to the Philippines and Japan, claiming thousands of lives and destroying homes.

Japan had had a tsunami alert system since the 1940s, and the United States had its own tsunami warning program based in Honolulu. But tsunamis know no borders — and in the aftermath of the Great Chile Quake, it became abundantly clear that the world needed an alert system that spanned the ocean basin. The new Pacific Tsunami Warning Center was responsible for sending alerts to 20 nations around the Pacific Ocean, based on data gleaned from several dozen seismic stations in the region, as well as tide gauges measuring sea level heights.

Four decades later, in 2004, the Pacific center was still the only basinwide tsunami monitoring system in the world. By then, the center had, in addition to the seismic and tide data, a handful of DART (Deep-ocean Assessment and Reporting of Tsunamis) buoys that tracked changes in seafloor pressure in real time — essential information to confirm, as quickly as possible, that a dangerous tsunami had actually been generated.

“These are no-notice events,” says Laura Kong, director of the International Tsunami Information Center, based in Honolulu. “We don’t know when they’re going to be, and we know that absolutely they can happen anywhere.” Although 70 percent of recorded tsunamis have happened in the Pacific Ocean, Kong says, tsunamis have occurred in ocean basins around the world.

An Indian Ocean tsunami blind spot

On December 26, 2004, a seismometer in Australia detected a large quake off the coast of northern Sumatra in the Indian Ocean (SN: 1/5/05). Researchers at the Pacific Tsunami Warning Center scrambled to determine the epicenter of the quake, and whether a tsunami had been generated. Within minutes, the center issued its first bulletin in response to the event, noting no tsunami threat to coastlines around the Pacific Ocean from this earthquake.

But it was possible — in fact, probable due to the magnitude of the quake — that a tsunami may have been generated in the Indian Ocean, near the quake’s epicenter, the center said. As reports from seismic stations around the world continuously upgraded the power of the quake off Sumatra, the probability grew that a huge, basinwide tsunami was already on the move.

The trouble was, without sea-surface height data from the region, there was no way to “see” it.

“The way we found out about the destructive nature of that tsunami was we were going through the internet,” says Stuart Weinstein, deputy director of the Pacific Tsunami Warning Center. “You’re looking for information on this big earthquake, and we weren’t able to get through to Indonesia or Thailand. There were no sea-level stations in Indonesia, nothing in Sri Lanka or India. We talked to our colleagues in Australia to see if they heard anything, and they said, ‘Nope.’”

It was a news story from Reuters, describing the devastation in Thailand, that first revealed to him just how destructive this tsunami was, he says. “That’s not the way a scientist [on duty] wants to find out.” He and other researchers began to reach out to embassies along Africa’s east coast to warn them that the waves were coming.

“That’s not the way a warning system should operate,” he says. “And the world learned a painful lesson: You can’t develop a warning system in a few hours.”

‘You have only 15 minutes’

In the aftermath of the Indian Ocean tsunami, tsunami preparedness became paramount.

It was “a powerful catalyst for change,” says Nelly Florida Riama, the deputy head of geophysics for Indonesia’s Meteorological, Climatological and Geophysical Agency, known as BMKG.

Before 2004, tsunami risk in the country was considered low, and there was no tsunami warning system, Riama said December 11 during a news conference at the American Geophysical Union annual meeting in Washington, D.C. Indonesia’s seismic stations were also capable of recording earthquakes only up to magnitude 6.5. “As the catastrophic events unfolded, it became very clear that the magnitude far exceeded this threshold,” she said.

The United Nations organized meetings to push for more preparedness, including in Indonesia, as well as in vulnerable island nations such as Samoa and Tonga. “They saw what happened, and they stepped up awareness, national drills and exercises,” Kong says. Governments devised evacuation maps, public service announcements, cell text notification systems.

UNESCO’s Intergovernmental Oceanographic Commission met in February 2009 in Apia, Samoa, to help raise public awareness that, in the worst-case scenario, villagers would get just 15 minutes of warning before a tsunami struck.

That was “our number one point made: You have only 15 minutes. Remember that number,” Kong says. “And that’s exactly what happened.”

Just seven months after the IOC meeting, two large earthquakes struck back-to-back on the northern Tonga trench, generating a series of tsunamis up to 22 meters tall that engulfed the coasts of Samoa, American Samoa and Tonga.

The 15-minute message, the constant drills and the heightened awareness of the warning signs — powerful ground shaking, suddenly retreating coastal waters — had taken root. Villagers had their plans to evacuate to higher ground in place. One school principal in Poloa, American Samoa, didn’t wait for an official warning before swiftly closing his school and evacuating his students along a coastal road to the mountains, just before the waves arrived.

Hundreds of people on the island still died. But without all those efforts, the loss of life in the islands would have been even more sobering, Kong says. “Preparedness is what saved so many lives.”

Speeding tsunami warnings

In 2004, researchers did manage to issue a “rudimental” tsunami forecast, says Vasily Titov, a tsunami scientist at the U.S. National Oceanographic and Atmospheric Administration’s Pacific Marine Environmental Laboratory, based in Seattle. “It was, quantitatively, very approximate. And it was very late.”

So late, in fact, that “it was difficult to call a forecast,” he adds. The data were assembled manually, and the forecast wasn’t completed until the wave was already exiting the Indian Ocean.

Timely forecasts, Titov says, required two things that weren’t readily available at the time: much faster and more accurate models, and data that could be fed into the models in real time.

Real-time data collection has improved dramatically since then: There are now 75 DART buoys spread around every ocean, covering every coastline. “In terms of technology, we can detect any tsunami from any major fault,” Titov says.

Better tsunami forecasts also require better understanding of the relationship between, for example, earthquake magnitude and tsunami size and energy. New technologies to trace the imprints of prehistoric and historic tsunamis in sediments and corals are providing additional clues to tsunami generation.

But the greatest advances are in observation and alerts. There are tsunami warning centers now monitoring most of the world’s ocean basins. These sea level observation systems have grown exponentially: In 2004, just one sea level station was monitoring the Indian Ocean. Today, there are some 1,400 stations delivering real-time sea-surface height data in that ocean basin, which aid in forecasts not just for tsunamis, but also for cyclone-related storm surges.

Faster supercomputers are aiding in speeding up warning systems, in hopes of adding a few more precious minutes for people to get to safety. Seismic analyses that took five to six minutes in 2004 now take just a minute or so. A new technique to determine what’s called the Centroid Moment Tensor, or CMT, allows researchers to quickly assess the geometry of faults that have slipped — and that makes it possible to determine how much the seafloor might have been pushed up or moved sideward, generating a tsunami.

Sea level assessments that in 2004 might have taken hours have dropped to about an hour or less — thanks not just to hundreds more stations measuring sea level, but also to the speed of data transmission. In 2004, most sea level stations transmitted data once an hour and sampled sea-surface heights every six minutes. Now it takes a fraction of the time to track swift changes in sea level, with stations measuring sea-surface heights every minute and uploading those data every five minutes.

To improve early warnings of tsunamis generated by nonseismic sources, the Pacific Tsunami Warning Center is working with scientists at the University of Hawaii to develop detection methods based on infrasound, sound waves of very low frequency that can be generated by breaking waves (SN: 6/25/18). Infrasound intensity has been found to be correlated with wave heights, making it a promising detection system, Weinstein says.

Researchers are also hoping to partner with communications companies installing transocean fiberoptic cables. Turning these communications cables into “smart” cables by instrumenting them with pressure sensors, accelerometers and other devices could greatly enhance tsunami detection around the globe.

Another burgeoning area of tsunami research focuses on how the walls of water might impact coastal buildings. After the 2011 earthquake and tsunami that struck northwestern Japan, damaging the Fukushima nuclear facility, engineers collected information on the damage to buildings: how the foundations and structural elements were impacted, and how the angle of impact — head-on, or from the side — altered those impacts (SN: 3/14/11). That information led to an international building code for tsunami-resistant structures.

A grim sort of tsunami success story

When it comes to reducing loss of life, public awareness campaigns have been the most significant advance of the last two decades, scientists say.

Japan’s 2011 event is a case in point for how preparedness can mitigate — but not prevent — disaster, Kong says. “Japan is arguably the best-prepared system in the world. But tsunamis don’t discriminate.” Some 18,000 people died as a result of that earthquake and tsunami, she notes. It’s a horrifying number — but it also represents just 5 percent of the people considered to have been vulnerable and in the tsunami’s path. The death toll might have been so much worse — perhaps in the hundreds of thousands, given the region’s population density.

The Hokkaido-Nansei-Oki earthquake of 1993 that spawned one of the largest tsunamis in Japanese history killed 15 percent of the people estimated to be in its path. A tsunami that swept across Papua New Guinea in 1998 killed 75 percent of the vulnerable population. And the 2004 tsunami killed 90 percent — or 130,000 people — in the heavily impacted parts of Banda Aceh, Indonesia.

If Indonesia had had a functioning monitoring system, thousands might have been saved there, though they still would have had little time to get to safety, Weinstein says. And, he notes, a working Indian Ocean tsunami warning system would have saved a majority of the lives lost in other nations farther from the epicenter.

“When you’re struck by a tsunami that’s like 30 meters tall, it’s going to be difficult to save everybody,” he says. “But you can save most.”

Looked at that way, 5 percent mortality is a grim sort of success story, a testament to the world’s growing awareness and readiness for the swift fury of the sea. Kong says: “We’ve all come a long way since 2004.”

But “we are not done yet,” Titov says. Since 2004, over 20,000 people have died in over 50 tsunamis. The waves move so quickly that most fatalities occur near the tsunami source — just as they did in 2004. Speeding up local tsunami forecasts is “the biggest challenge. We need quicker detection, and quicker models.”

In November, an international group of tsunami researchers and policy makers met in Banda Aceh to reflect on the last two decades of progress. From that meeting, hosted by UNESCO, the general consensus emerged to set an ambitious goal of 100 percent tsunami readiness in all at-risk communities around the world by 2030.

That will require not just improved technology, but also education and planning, Titov says. “It’s a huge challenge, but it’s what we’re striving for.”