Sediment cores pulled from a stretch of Alaska’s Aleutian Islands hold evidence of eight separate tsunamis striking Driftwood Bay over the past 2,000 years, an average of roughly one every 250 years, according to research the U.S. Geological Survey published from work along the Alaska-Aleutian megathrust. The fault is one of the largest and most active subduction zones on the planet, the boundary where the Pacific Plate grinds beneath North America, and it has already produced two of the twentieth century’s most destructive Pacific-wide tsunamis, in 1946 and 1957.
The history here matters well beyond the Aleutians themselves. A megathrust rupture along this boundary sends waves not just onto Alaska’s own coastline but across the open Pacific toward Hawaii, the West Coast and beyond, and the 1957 event reached Hawaii with enough force that researchers are still refining exactly how it happened almost seventy years later.
Eight Sand Sheets at Driftwood and Stardust Bays
The evidence takes the form of buried sand. Geologists digging into the marshes behind Driftwood Bay found eight distinct sheets of landward-fining marine sand, some extending as far as 375 meters inland and 23 meters above the mean tide line — a pattern consistent with a 164-to-257-year recurrence interval for large eastern Aleutian tsunamis, each sheet a signature of a tsunami large enough to carry ocean sediment deep into the coastal lowland. The study, led by USGS geologist Rob Witter and published in the Geological Society of America Bulletin, dated the layers using sand-sheet stratigraphy, stranded drift logs, boulder distribution and volcanic ash layers that serve as time markers between sites.
A second site, Stardust Bay, sits roughly 200 kilometers, or 124 miles, from Driftwood Bay, and its sediment record largely agrees with what researchers found at the first location. Two sites that far apart recording a similar sequence of ancient tsunamis makes it harder to explain the pattern as a local quirk of one bay’s geography.
A Recurrence Interval of 164 to 257 Years
Run the arithmetic on eight events across two millennia and the average gap comes out close to 250 years, squarely inside that measured range. It is a range built from dating uncertainty in the sediment layers, not a claim that the fault fires on a fixed clock — the true interval could sit anywhere from a century and a half to two and a half centuries between events. Eighty-eight years have passed since 1957.
Witter’s team set out to test something specific: whether the “creeping” sections of the megathrust, where the plates slide past each other more gradually, posed a lower earthquake and tsunami hazard than the “locked” sections that build up strain for centuries before rupturing violently. The sediment record did not support that assumption. Tsunamis traced to both types of fault segment showed up in the same sand layers, at comparable scale.
The 1957 Earthquake That Still Anchors the Record
Every recurrence estimate here is calibrated against one known event. The 1957 earthquake registered magnitude 8.6 and ruptured roughly 1,200 kilometers of the plate boundary, generating tsunami waves between 6 and 12 meters — 20 to 40 feet — high that struck both the Aleutians and Hawaii. A closer look at the rupture found shallow slip of 12 to 26 meters across a 600-kilometer eastern section alone, a patch large enough on its own to equal a magnitude 8.3 to 8.6 earthquake and to explain Hawaiian tsunami runup that older models had struggled to account for.
The 1946 earthquake nearby produced a similarly destructive tsunami across the Pacific, and comparing the two ruptures is part of how USGS keeps refining what a “typical” large event on this fault actually looks like versus what a worst-case one could do.
Rich Briggs: A More Complicated Tale Than Modern Instruments Tell
USGS scientist Rich Briggs summed up what the longer record adds: “The long tsunami archive, written in the geologic record at Driftwood and Stardust bays, is in good agreement, and it tells a more complicated tale than the simple story.” Modern seismic instruments have only been running for a few decades, far too short a window to capture a fault that ruptures on a 164-to-257-year cycle even once, let alone often enough to characterize its range of behavior.
Closing that gap is the point of USGS’s ongoing paleoseismology work in the Aleutian, Kodiak and Semidi islands, research that feeds directly into the National Seismic Hazard Maps and NOAA’s tsunami inundation models used up and down the West Coast. Field teams there study tectonic tremor and slow-slip events alongside the older sediment layers, looking for any sign of how a locked or creeping segment behaves in the years before it ruptures rather than only after.
Eighty-eight years into a recurrence window that could run anywhere from roughly a century and a half to two and a half centuries, the sediment record buried at Driftwood and Stardust bays remains the only instrument that has been running long enough to say what comes next, and how far its warning would travel once it did.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
