The Repository Hook Echo

Alaska 1964 Earthquake: The Coastline That Rose 38 Feet

October 15, 2026 · Runtime 15:08 · Watch on YouTube

On Good Friday, 27 March 1964, a magnitude 9.2 earthquake, the largest ever recorded in the United States, shook south-central Alaska for as long as four and a half minutes. Along hundreds of miles of coast the land itself moved: some shores sank as much as 8 feet, others rose as much as 38. The nearest seismograph went blind and the tide gauges in the zone were destroyed, so the movement was measured with barnacles. Federal film from 1964 to 1966 shows Anchorage, the waterfronts and the shores.

Transcript

Anchorage, Good Friday 1964 0:00

This is Anchorage, Alaska, a few days after Good Friday, 1964. The ground here shook for as long as four and a half minutes. It was a magnitude nine point two, the largest earthquake ever recorded in the United States.

A coastline that moved 0:16

But the strangest thing it did was not in any city. Along hundreds of miles of coast, the land itself moved. Some shores dropped by as much as eight feet. Others rose out of the sea by as much as thirty-eight. And almost no instrument saw it happen.

The nearest seismograph went blind. The tide gauges were destroyed or washed away. So how do we know how far a coastline moved? The answer came from barnacles. And what they measured forced geology to accept a new idea about how the Earth works.

5:36 p.m., Prince William Sound 0:51

The earthquake struck at five thirty-six in the evening, on Friday, March 27th, 1964. The epicenter was in northern Prince William Sound, about ninety miles east of Anchorage. Below it, a section of fault about five hundred and eighty miles long began to break.

On the clocks of the world's observatories, it was three thirty-six on Saturday morning, March 28th, Greenwich time. Sitka had been a seismological observatory since 1904. It was about to record the largest earthquake of its life, and so was every seismograph on Earth.

Four and a half minutes 1:27

Most earthquakes last seconds. This one did not. An instrument at a power plant in Whittier, on bedrock, recorded vibrations for nearly four minutes. People who timed it with their watches reported between two and a half and five minutes.

Turnagain Heights 1:44

In Anchorage, the shaking found a weak layer in the ground, a clay laid down in an old glacial estuary, called the Bootlegger Cove Clay. It holds a lot of water, and in places it is sensitive: disturb it, and it loses most of its strength. Under minutes of vibration, that is what happened. At Turnagain Heights, a bluff above the water slid toward the sea.

About a hundred and thirty acres were destroyed. About seventy-five homes went with them. The land along the bluff, more than a mile and a half of it, dropped an average of about thirty-five feet and broke into blocks.

Fourth Avenue 2:20

Downtown, the same clay moved under Fourth Avenue. All or part of fourteen city blocks slid sideways. Where the ground pulled apart, a strip of street dropped eleven feet, and the storefronts on one side ended up below the sidewalk on the other.

Valdez and Seward 2:36

But the deadliest moments were on the waterfronts. At Valdez and at Seward, the towns were built on loose sediment at the edge of deep fjords. The shaking turned that sediment to slurry, and the waterfronts slid into the sea.

At Valdez, the slide took the docks and raised waves that left marks as high as a hundred and seventy feet above the fjord. Thirty people died there. Geologists later recommended moving the whole town, and it was rebuilt about four miles away.

At Seward, a strip of the shore about four thousand feet long slid into Resurrection Bay, with its docks and railroad wharves. Oil tanks along the shore ruptured and caught fire. Then the waves came in over what was left.

The Geological Survey counts a hundred and twenty-nine deaths in all. In southern Alaska, seventy percent of them were caused by waves from underwater landslides like these, not by the shaking itself. The tsunami also crossed the Pacific, and killed sixteen people in Oregon and California.

The instruments of 1964 3:39

Now, the instruments. In 1964 there were only two seismic stations operating in Alaska, at Sitka and Fairbanks. And according to the Geological Survey, there were no instruments in Alaska capable of recording this earthquake.

At College, near Fairbanks, the seismograph wrote with a beam of light on photographic paper. When the waves arrived, the beam moved so fast that the paper never caught it. Then the galvanometer mirror stuck at the edge of its range, and stayed there until someone reset it the next day.

Far from Alaska, the instruments kept writing for hours. On a record from Weston, Massachusetts, where each line is thirty minutes long, the earthquake fills line after line. At Dallas, the light beam of one seismograph was lost for many hours after the waves arrived.

8.4, and why it was wrong 4:30

Far away, the instruments did record it, and the magnitude they produced was eight point four. That number was wrong, and not because anyone made a mistake. The scale itself had a ceiling.

Pasadena read eight point four. The Coast and Geodetic Survey, eight point three to eight point four. Berkeley, up to eight point seven five. Every observatory agreed it was enormous. And every one of them was reading a scale that could not go much higher.

The magnitude of that era was read from surface waves a few tens of kilometers long. When a fault breaks over hundreds of kilometers, those waves stop growing. The scale saturates. In 1977, Hiroo Kanamori computed it from the seismic moment instead, a measure of the whole rupture. The answer was nine point two.

The difference between eight point four and nine point two sounds small. On the magnitude scale, every step of one is about thirty-two times more energy. Point eight of a step is about sixteen times. The first number missed most of the earthquake.

The tide gauges that failed 5:40

Then there were the tide gauges, the one kind of instrument that measures the sea against the land. Three of them sat in the zone that moved, at Seward, Kodiak and Homer. Every one of them failed.

The Seward gauge went into the bay with the submarine landslide. The Kodiak gauge was washed away half an hour after the earthquake, with its record for the month of March. The Homer gauge was knocked out by the shaking.

One of the Survey's reports put it in a single sentence. Instrumental records of the time and rate of the movement, in the area that moved, are nonexistent. Hundreds of miles of coast rose and sank, and no machine was there to watch. Whatever measured it had to be something that was on the shore before the earthquake, and was still there after.

George Plafker and the barnacles 6:28

So a geologist went to look. George Plafker, of the Geological Survey, was part of the first team of earth scientists to reach the region. He went looking for something that every rocky shore in Alaska carries without anyone writing it down.

On a rocky coast, barnacles and a brown seaweed called Fucus grow up to a fixed height above the low tide, and no higher. Above that line they dry out and die. Their upper limit is a tide mark, drawn by living things, on every rocky shore in the Sound.

Where the land had risen, that band was now high and dry, out of reach of the tide. Where the land had sunk, it was underwater. And by July of 1964, a new line of young barnacles had settled at the new water level.

A level and a rod 7:15

The distance between the old line and the new one was the movement. Plafker measured it with a hand level and a surveyor's rod. He calibrated the method at seventeen places where tide records gave the answer, and judged it good to within a foot, and in most places half a foot.

There was a second check, on land. The Coast and Geodetic Survey re-ran seven hundred and twenty-two miles of first-order leveling lines that had been surveyed before the earthquake. Benchmark by benchmark, the land could be compared before and after.

38 feet up, 7.5 feet down 7:47

Here is what the barnacles said. On Montague Island, in a narrow belt less than ten miles wide, the land rose thirty-eight feet. On the southwest coast of the Kenai Peninsula, it sank seven and a half.

On Montague Island, the earthquake also broke the surface. Faults branching up from the main one reached daylight, and the Patton Bay fault shows a vertical offset of twenty to twenty-three feet. Along those faults, the Survey mapped uplifts of thirty-six feet.

Middleton Island, out in the Gulf, came up about ten feet, and a new terrace of beach appeared around it. Portage sank about five and a half feet. Anchorage, about two. A geologist had measured the deformation of a coastline, one shore at a time.

The number grew as more shores were measured. In his first paper, in 1965, Plafker reported as much as ten meters of uplift on land. By the final report, in 1969, the measured maximum was thirty-eight feet, about eleven and a half meters.

Our measurement: the Seward gauge 8:53

There is one place where we can check those numbers against a machine. Seward's tide gauge had been recording since well before the earthquake, and after it was destroyed, a new gauge picked up the record. We pulled every monthly average from the federal archive.

From 1959 to 1963, the average sea level at Seward sat at one height. From 1965 to 1969, it sat three point four feet higher. The sea had not risen. The town had sunk. The Survey's own leveling put Seward's drop at about three and a half feet.

And the gauges outside the zone are the control. Sitka, Yakutat, Juneau and Ketchikan, all recording before and after, moved by a few inches at most. The largest change among them was Juneau, about three inches, and in the other direction. The step at Seward belongs to the earthquake.

Portage and the sea 9:50

Where the land sank, the sea came in to stay. Along Turnagain Arm, near Girdwood and Portage, coastal forests and marshes dropped below the high tide, and the salt water killed them. Girdwood, a hamlet of eighty people on the shore, had to be moved.

Two belts 10:07

Put all the measurements on one map, and a shape appears. Vertical change covered more than a hundred and ten thousand square miles, about two-thirds the size of California. And it came in two long belts, side by side.

Inside the sinking belt, about forty-eight thousand square miles, the land dropped an average of roughly two and a half feet. Across broad parts of the rising belt, it came up about six. A whole region tilted, in a few minutes.

The belt nearest the ocean went up, Montague Island, Middleton Island, the outer coast. The belt farther inland went down, the Kenai Peninsula, Kodiak Island, Anchorage. Up on the side of the sea, down on the side of the land.

In 1964, geology had no good explanation for that. The Survey's own account says that at first, geologists did not know how such a huge earthquake could have happened, because the theories of the day could not explain so large a movement. Plafker's map could.

The megathrust 11:10

The only fault that fits both belts is a gently tilted one, running under the continent from the deep trench offshore. Plafker called it a megathrust. From the aftershocks, he put the part that moved at five hundred and fifty to six hundred miles long, and over a hundred miles wide.

Below that fault, the floor of the Pacific is sliding under Alaska. For centuries the two plates stay locked. The edge of the continent is dragged down and squeezed, and the land behind it bulges. Then the lock breaks, and everything springs back at once.

The edge near the ocean snaps up. The land behind it drops. That is exactly the pattern in the barnacles. Current estimates put the average slip on the fault at around thirty feet, and some patches slipped as much as sixty.

The slip was not even. A 1996 study using the tsunami and the survey data found two patches that carried most of it. One under Prince William Sound slipped about eighteen meters on average. One near Kodiak, about ten.

Evidence for plate tectonics 12:16

Plate tectonics was still a young idea in 1964. The Survey calls this the first earthquake generally understood as having happened on a subduction zone, the boundary where one plate dives under another. The evidence was a band of seaweed and shells.

The fault kept moving for months. In the first day alone, there were eleven aftershocks larger than magnitude six. Thousands more followed, and their positions drew the outline of the rupture that Plafker's shores had already sketched.

The tsunami warning center 12:47

In 1964, Alaska's tsunami warnings came from Hawaii. In 1967, a tsunami observatory opened at Palmer, Alaska, as a direct result of this earthquake. Today it is the National Tsunami Warning Center.

The coast's own record 13:03

And the coast had done this before. Middleton Island carries five older beach terraces above the one made in 1964, each one lifted by an earlier earthquake. Driftwood on them gives repeat times from four hundred to twelve hundred years, eight hundred on average.

In the Copper River Delta, cores through layers of buried marsh and forest record nine great earthquakes before 1964, over five thousand six hundred years. That is about one every six hundred years. The coast keeps its own record, in mud, wood and shells.

Studies at Turnagain Arm and Kodiak put the average interval at about five hundred and thirty-five years. And they suggest that the last earthquake like 1964 may have struck about eight hundred and eighty years ago. Each estimate comes from a different place, and they all say centuries, not decades.

Still moving 13:57

It is also still moving. Seward's tide gauge has stood on the same pier since 1966. Since then, the sea there has dropped about six inches against the land, while the world's oceans rose. The ground the town sank on is slowly coming back up.

What the record supports 14:14

Here is what the record supports. The seismographs were blinded or saturated, and the tide gauges were destroyed. The best instrument for the largest earthquake in American history was the shoreline itself, read by a geologist with a level and a rod. One tide gauge that survived in the record agrees with him to within a few inches. And the pattern he drew is now the textbook picture of a subduction earthquake, from Alaska to Chile to Japan.

Next: Harvey, 2017 14:41

Next, a storm whose number came from one simple instrument. In August 2017, Hurricane Harvey stalled over southeast Texas. A rain gauge near Nederland collected sixty point five eight inches, the most significant tropical cyclone rainfall event in United States history. That is the next episode. Harvey, 2017: sixty inches of rain.

Description and sources

The instruments of 1964 did not capture this earthquake. At College, Alaska, the seismograph's light beam moved too fast to mark the paper and its mirror jammed until the next day; the surface-wave magnitude of 8.4 had a ceiling, and Kanamori's seismic moment (1977) put it at 9.2. All three tide gauges in the deformed zone failed. USGS geologist George Plafker measured the movement from the upper growth limit of barnacles and Fucus seaweed: up to 38 ft of uplift on Montague Island, up to 7.5 ft of subsidence on the Kenai Peninsula. Up on the seaward side, down inland: only a megathrust fits, the first earthquake generally understood as a subduction-zone rupture. Our measurement of NOAA's Seward tide gauge: mean sea level 3.4 ft higher in 1965-69 than in 1959-63 (USGS leveling: about 3.5 ft), against a few inches outside the zone.

PRIMARY SOURCES

USGS Fact Sheet 2014-3018 (Brocher et al.), The 1964 Great Alaska Earthquake and Tsunamis. https://pubs.usgs.gov/fs/2014/3018/pdf/fs2014-3018.pdf
Plafker (1969), USGS Professional Paper 543-I. https://pubs.usgs.gov/pp/0543i/pp543i_text.pdf
Plafker (1965), Science 148: 1675. https://doi.org/10.1126/science.148.3678.1675
Kanamori (1977), J. Geophys. Res. 82: 2981. https://authors.library.caltech.edu/records/rrk02-vn009
Hansen et al. (1966), USGS PP 541. https://pubs.usgs.gov/pp/0541/pp541.pdf
USGS PP 542-A (Anchorage), 542-C (Valdez), 542-E (Seward). https://pubs.usgs.gov/pp/0542a/
USGS, 1964 Alaska Earthquake Seismograms. https://earthquake.usgs.gov/earthquakes/statistics/events/alaska1964/1964seismogram.php
Johnson et al. (1996), J. Geophys. Res. 101: 523. https://doi.org/10.1029/95JB02806
Alaska Earthquake Center, aftershocks. https://earthquake.alaska.edu/event/96441lal4e/aftershocks
NOAA NTWC history. https://ntwc.ncep.noaa.gov/php/content/history.php
NOAA CO-OPS monthly means (Seward 9455090 and four control gauges); PSMSL station 266. https://tidesandcurrents.noaa.gov/
NHC, Hurricane Harvey report (next episode). https://www.nhc.noaa.gov/data/tcr/AL092017_Harvey.pdf

ARCHIVE PLATES — public-domain documents, with author, date, archive and licence.

2005-07-08 20:37:40 | Wikimedia Commons | Public domain
USGS — 2003-06-06 09:33:14 | Wikimedia Commons | Public domain
G Plafker — 1964-07-27 | Wikimedia Commons | Public domain
U.S. Geological Survey — 2018-11-30 23:50:23 | Wikimedia Commons | Public domain
Coast Geodetic Survey — 1964 | Wikimedia Commons | Public domain
U.S. Department of Defense — 1964 | Wikimedia Commons | Public domain

Hook Echo reads a disaster through the instruments that recorded it: radar, sensor, survey and report, with the primary sources above.

About the material. Video: three federal films on the Internet Archive, shown in black and white: "The Alaskan Earthquake, 1964" (produced for the U.S. Geological Survey, 1966; National Archives 46347; no copyright notice on the film), "The Big Picture: Alaskan Earthquake" (U.S. Army, 1966) and "The Face of Disaster" (U.S. Department of Health, Education, and Welfare; National Archives 61062). The closing aerial of Houston is U.S. Army Corps of Engineers footage from August 2017, released through DVIDS. Photographs: U.S. Geological Survey (including G. Plafker), Coast and Geodetic Survey and U.S. Navy, public domain via Wikimedia Commons. The tide-gauge series are our own measurement from NOAA's public data. Every diagram is drawn in code from the sources listed. Nothing in this video is AI-generated imagery.

#HookEcho #AlaskaEarthquake #Earthquake