Alaska Megatsunami Confirmed as Second Largest Ever Recorded

At 5:26 a.m. on August 10, 2025, an entire mountainside at the mouth of Alaska's receding South Sawyer Glacier detached and dropped into Tracy Arm fjord, approximately 50 miles south of Juneau. The collapse — more than 64 million cubic meters of rock, roughly a quarter-mile cube of earth — displaced enough water to send a breaking wave over 100 meters (300 feet) high hurtling down the narrow fjord at speeds exceeding 150 mph [1][2]. When the surge hit the opposite wall, water climbed 481 meters (1,578 feet) above sea level — 1.5 times the height of the Eiffel Tower [3].

A study published in Science in May 2026 confirmed the event as the second-largest tsunami ever recorded, behind only the 1958 Lituya Bay megatsunami at 524 meters (1,719 feet) [1]. Both events occurred in Alaska. Both were triggered by landslides. And both, by narrow margins of timing and geography, avoided killing hundreds or thousands of people.

Source: Science / USGS

Data as of May 8, 2026CSV

What Happened in Tracy Arm

Tracy Arm is a narrow, steep-walled fjord in the Tongass National Forest, one of Southeast Alaska's most popular destinations for cruise ships and tour boats. During peak summer season, the fjord sees up to 20 vessels daily, including six cruise ships carrying as many as 6,000 passengers each [4]. In total, Tracy Arm receives more than 500,000 visitors annually [5].

The collapse registered globally as a magnitude 5.4 seismic event [2]. The landslide stripped more than 900 feet of coastline into the water [6]. Inside the fjord's confined geometry, the initial wave created a standing wave — a seiche — that oscillated with a period of roughly 66 seconds, continuing for approximately one full day [1]. A long-period seismic signal generated by the event persisted globally for up to 36 hours [3].

The wave scoured mountain surfaces to bare rock as it traveled toward Tracy Arm's mouth [7]. The yacht Blackwood, a 103-foot vessel, reported significant rolling and rushing water. Christine Smith and her husband, captaining the 65-foot wooden vessel David B, had been forced by bad weather to anchor 50 miles from their planned location the night before. Had they anchored as planned, they would have faced what Smith described as "a deadly 300-foot wall of water heading their way" [5].

No one was killed. No one was injured. The event occurred at 5:26 a.m. on a rainy morning, when vessel traffic had not yet entered the fjord.

"This could be really catastrophic if it were to happen at 10 a.m.," said Ezgi Karasozen, a research seismologist at the Alaska Earthquake Center [8].

The Glacier Connection

The South Sawyer Glacier had been retreating for decades, but the final destabilization happened in days. Satellite data shows the glacier retreated approximately 500 meters in spring 2025, with much of the critical exposure occurring between August 2 and August 5 — just five days before the mountainside failed [1][9].

The mechanism is called debuttressing: when a glacier retreats, it removes the physical support that had been holding adjacent valley walls in place for centuries. The rock, no longer braced, becomes unstable.

Dan Shugar, a geomorphologist at the University of Calgary and lead author of the Science study, described it this way: "As that glacier retreated... it's like if you have a kid and they said they cleaned their room but really all they did was throw everything in the closet" [9]. The glacier's retreat exposed the structural weakness that had been hidden behind ice.

Bretwood Higman, a geologist and executive director of Ground Truth Alaska, confirmed: "That retreat was exactly at this place where the failure happened" [10]. He also acknowledged a gap in scientific monitoring: "We clearly did not identify this beforehand" [10].

Source: USGS / Satellite Data

Data as of May 8, 2026CSV

Glacier-covered areas across Alaska decreased 13% between 1985 and 2020 [9]. Leigh Stearns, a glaciologist at the University of Pennsylvania, noted that steep slopes respond sensitively to permafrost loss, glacier retreat, and increased soil moisture — all conditions that are intensifying as temperatures rise [11].

A Pattern of Acceleration

Tracy Arm is not an isolated event. It is part of a documented acceleration in landslide-generated megatsunamis in Alaska and comparable fjord environments worldwide.

Source: Science / Ground Truth Alaska

Data as of May 8, 2026CSV

Historically, these events occurred approximately once every 20 years in Alaska's fjord systems. In the decade from 2010 to 2019, two were documented. In the current decade, the count has already reached six — roughly a tenfold increase over the historical baseline [1][10].

The 2015 Taan Fiord event in Wrangell-St. Elias National Park foreshadowed Tracy Arm. There, 180 million tons of rock dropped into a fjord that had been filled by the Tyndall Glacier as recently as 1961. The glacier's retreat — 17 km of terminus withdrawal and more than 400 meters of ice thinning between 1961 and 1991 — had progressively destabilized the surrounding slopes. The resulting wave reached 193 meters (633 feet), the fourth-highest tsunami runup ever recorded [12][13].

InSAR (Interferometric Synthetic Aperture Radar) satellite imagery now reveals that many Alaskan slopes are actively moving, suggesting the current inventory of hazards is far from complete [1].

The Warning That Almost Worked

Retrospective analysis of seismic data from stations 60 miles from the Tracy Arm landslide revealed a sequence of precursor signals that, had they been recognized in real time, could have provided hours of warning [8][14].

At least 24 hours before the collapse, instruments recorded repetitive seismic activity resembling small earthquakes. Six hours before failure, the seismic intensity began increasing exponentially. In the final phase, the discrete signals merged into a continuous tremor — a "hum" of seismic energy that indicated the mountainside was in the process of failing [8].

Karasozen's team at the Alaska Earthquake Center is now developing landslide-detection algorithms that could recognize these patterns automatically. She and her colleagues have proposed a color-coded warning system: yellow alerts would trigger when rapid glacier retreat combines with heavy rainfall; orange when seismic precursors are confirmed; red when failure is imminent [8][14].

Even after onset, the system could provide approximately 10 minutes of warning for distant vessels — enough time, in theory, to move away from fjord walls and into deeper water [14].

But no such landslide monitoring system currently operates at this scale anywhere in the United States [8].

"We should be acting more quickly," Karasozen said [8].

Why No Warning System Exists

The 1958 Lituya Bay megatsunami — which sent water 1,719 feet up a mountainside, killed two people, and was witnessed by only six — established landslide-generated megatsunamis as a known hazard in Alaska more than 65 years ago [15][16]. The question of why no dedicated monitoring network has been built since then involves overlapping jurisdictional, funding, and scientific obstacles.

The existing U.S. tsunami warning infrastructure, operated by NOAA's National Tsunami Warning Center in Palmer, Alaska, is designed primarily for earthquake-generated tsunamis in the open ocean — waves that travel across ocean basins over hours. Landslide-generated megatsunamis in fjords are a fundamentally different problem: they occur in confined waters, they generate and dissipate locally within minutes, and they are not detectable by the deep-ocean DART buoy network [8][14].

Building a fjord-specific monitoring network would require geophysical sensors (seismometers, InSAR satellite coverage), meteorological stations, rapid communication infrastructure, and coordination across multiple federal agencies (USGS, NOAA, the Forest Service, the National Park Service) and the state of Alaska [14]. The Barry Arm landslide in Prince William Sound — a 500-million-cubic-meter mass that could threaten the town of Whittier — is one of the few sites that has received dedicated instrumentation from USGS and Alaska's Division of Geological and Geophysical Surveys [17][18]. But Barry Arm's current movement rates are slow, and revised modeling suggests a resulting wave would reach only about 2 meters at Whittier, significantly less than earlier estimates [18].

The fundamental tension is between risk probability and resource allocation. Both of Alaska's record megatsunamis occurred in remote, sparsely populated areas. Tracy Arm's fjord had no permanent residents; Lituya Bay had three fishing boats. Critics of large-scale monitoring investment can point to this record and argue that the actual toll — two deaths in 67 years — does not justify a multi-agency monitoring buildout.

Researchers counter that the risk profile is changing. Tracy Arm hosts half a million visitors annually. Cruise tourism in Alaska carried 1.6 million passengers the year before the event [5][4]. The intersection of accelerating slope instability and growing human presence in fjord environments is creating conditions that did not exist when Lituya Bay set its record.

Global Comparisons

Alaska is not the only region facing this hazard, but it holds a disproportionate share of the record.

Greenland's Karrat Fjord (2017): A massive landslide on the west coast of Greenland sent water more than 90 meters up the adjacent slope and 50 meters up the opposite hillside across the 6-km-wide fjord. The wave reached the nearby settlement of Nuugaatsiaq, destroying 11 homes and killing four people [19]. Denmark and Greenland's geological survey subsequently increased monitoring in the region.

Canary Islands / Cumbre Vieja: A 2001 research paper proposed that a volcanic flank collapse on La Palma could trigger a trans-Atlantic megatsunami reaching the U.S. East Coast. The scenario received widespread media coverage and was dramatized by Netflix. However, subsequent peer-reviewed research has largely debunked the extreme scenario. No geological evidence of any past megatsunami from the Canary Islands has been found, and revised worst-case modeling suggests waves reaching the U.S. coast at 6-7 feet — significant, but not civilization-ending. Over 17 eruptions have occurred in the Canary Islands since the 1400s without producing a megatsunami [20][21].

Norway: Norwegian fjords face similar landslide-tsunami risks, and Norway's geological survey (NGU) operates an active monitoring program for unstable rock slopes, including the Åknes rock slope in Storfjord, which has continuous real-time monitoring and an automated warning system linked to local evacuation plans. This system is arguably the most advanced landslide-tsunami early-warning network in the world [22].

The contrast with the U.S. is stark. Norway monitors a known unstable slope in a fjord with a few thousand residents downstream. Alaska has dozens of potentially unstable slopes in fjords that collectively host hundreds of thousands of seasonal visitors, and no comparable monitoring system.

Source: OpenAlex

Data as of Jan 1, 2026CSV

Liability, Land Use, and the Aftermath

Tracy Arm lies within the Tongass National Forest, managed by the U.S. Forest Service. Tour operators, cruise lines, and fishing vessels operate in the fjord under various federal and state permits. The question of who bears liability for land-use decisions and tourism permits in identified high-risk zones has no clear answer in current law [5].

At least six cruise lines, including Carnival Corporation, altered their 2026 Alaska itineraries in response to the Tracy Arm event, removing or modifying fjord excursions [23]. This is an industry self-regulation response, not a government mandate.

As of May 2026, no federal or state agency has publicly updated hazard maps or permitting rules for Tracy Arm or comparable fjord environments in Southeast Alaska. The USGS has published its analysis of the event and the Science paper has entered the public record, but the translation of scientific findings into regulatory action remains incomplete [1][2].

NASA's SWOT (Surface Water Ocean Topography) satellite, which happened to capture observational data from the Tracy Arm event, provided researchers with the first direct satellite observation of a megatsunami in progress — a dataset that Thomas Monahan, a researcher at the University of Oxford, described as unprecedented: "Until now, we simply didn't have a way to observe these waves directly" [3].

What Comes Next

The scientific community is clear on the trajectory. Glacial retreat is accelerating. Permafrost is degrading. The slopes that glaciers once held in place are becoming unstable at rates that exceed historical baselines. The number of landslide-generated megatsunamis in Alaska's fjord systems has increased roughly tenfold in a single decade [1][10].

The question is whether this acceleration will be met with a proportional response in monitoring, warning systems, and land-use regulation — or whether the next event will test the same gap between known hazard and institutional action that Tracy Arm exposed.

Michael E. West, director of the Alaska Earthquake Center and state seismologist, has noted that the seismic precursor data from Tracy Arm provides a concrete foundation for building a warning system [14]. Karasozen's team is testing landslide-detection algorithms against the Tracy Arm data [8]. Shugar and his co-authors have published their methods and measurements as a baseline for future monitoring [1].

The infrastructure to act on that science does not yet exist. Tracy Arm's 1,578-foot wave left no human casualties. The next one may not arrive at 5:26 on a rainy morning.