Every year around this time, NOAA, in concert with local, state, federal, and international organizations conducts a series of disaster drills. These drills are to raise awareness of tsunami and they give emergency management practice in what to do in the event of one along a coastline.
Tsunami are rare enough that to an extent they are an underappreciated hazard in the United States. After the Indian Ocean Earthquake and Tsunami of 2004, a need was recognized for warning systems in seas that previously were thought immune. The drills are done for both the Atlantic and Pacific, despite the rarity of tsunami in the Atlantic Ocean.
Throughout the post I’ll be using the term “M” for magnitude. I’m universally, unless otherwise indicated, referring to “moment-magnitude” and not the catchall “Richter Scale”. This is also denoted as “Mw”. Moment-magnitude is the preferred scale for large earthquakes for a variety of reasons.
Overflight of Sendai, Japan, after the March 11, 2011 M9 Tohoku Earthquake drowned some 20,000 people.
Tsunami (Japanese: “Harbor Wave”), are large waves caused by the displacement of water in an ocean or large lake. Earthquakes, volcanoes, landslides both above and below the water, other disturbances and explosions beneath the sea, and meteor impacts can cause tsunami. These waves are different than storm-driven waves as they have a long amplitude from the top of the wave (the crest) and the bottom of the wave (the trough). When they reach land, they often do resemble a tidal surge, thus the reason they are often called “tidal waves.”
There are also meteo-tsunami, which are caused by sudden pressure differences and air pressure disturbances. A line of intense thunderstorms can cause them. They are distinct from ordinary storm-surges and are actually quite common.
The United States can be affected by tsunami from the following sources and causes, all of which we’ll look at in detail.
Tsunami Hazard Assessment for US and US Territories
Earthquakes: For the most part, the largest and most far-reaching tsunami are caused by earthquakes along subduction zones. These zones, where one tectonic plate overrides another and forces it downward beneath it into the mantle, produce great thrust quakes that displace the seabed, and the ocean above it. Prior to 2004 it was believed that not every subduction zone was capable of creating great earthquakes (the ones exceeding M8) based on age of the crust being subducted and so on. We now know this isn’t exactly the case.
The Aleutian Arc and Cascadia are two subduction zones off of the West Coast (near-field), but the West Coast and Hawaii can be struck by waves generated from earthquakes off of Japan, Russia, the Philippines, Polynesia, and Central and South America (far-field). The Atlantic is a narrower ocean, and has very few subduction zones, none of which are well understood. The boundary between the Caribbean and North America is one, and another exists in the South Atlantic south of the Falkland Islands. An ill-defined boundary exists where Africa and Europe meet at Gibraltar; this area is the source for the M9 1755 Lisbon earthquake that birthed modern seismology.
Earthquakes can also trigger underwater landslides. These slides also displace the ocean. Because they occur so close to shore, they can be a significant hazard. These underwater landslides can also happen in lakes, although I personally haven’t learned of any research that suggests they can in the Great Lakes. It has happened before, in Lake Geneva at least. Sediments can also become unstable for a variety of reasons and collapse even without a triggering earthquake.
Volcanoes: Volcanoes can too cause tsunami, either through utterly exploding (like Krakatoa) or through their eruptive processes (pyroclastic flows reaching the sea).
Finally, asteroids. There’s an interesting, albeit unproven hypothesis that sometime shortly after the end of the last ice age, the Earth moved through a fairly dense field of debris, causing several bolide strikes and contributing to the “Flood Myths” that creationists point to as universal evidence of a fully-globe covering flood. They’re wrong, of course, very very wrong, and the hypothesis has little support. But, an asteroid dropping into one of the oceans would create massive tsunami.
Let’s take a look at the mechanics of subduction.
Subduction
One tectonic plate meets another. It forces itself up and over the other plate, forcing it down into the mantle. The action creates mountains, volcanoes, from where the down going plate melts, and earthquakes. Where a continent meets an ocean, you get a trench. Where two continents meet, you get the Himalayas. Where two oceans meet, you get a volcanic island arc.
The plates themselves are driven by currents within the mantle. Think of us as living on a thin crust which floats on molten melted glop. Everything is in motion. The Atlantic is roughly 20 feet wider today than it was when Columbus first crossed it.
The crust gets recycled. Running the length of the Atlantic is a distinct series of mountain ranges. Here, crust is upwelled from deep within the Earth, pushing North America away from Europe. It’s not an even process, of course, and the Mid-Atlantic Ridge is offset by fracture zones. And then it gets subducted at the other end, the surface expression of which is often a deep trench, diving down into the mantle to get burped up again at another fracture zone and oceanic ridge.
Sometimes, the process of subduction gets stuck. The two plates continue to converge, but a significant portion of their meeting place—the interface---is not moving. Pressure builds. On the surface you’ll see the crust deform in subtle but recognizable ways. Beaches get uplifted. And so on.
Eventually something’s gotta give—and we get an earthquake---and everything jumps. If the quake is big enough the seabed is ruptured. Land sinks and rises significantly, and a tsunami is born.
The United States has three subduction zones within its waters, two in the Pacific and one in the Atlantic. Let’s take a close look at all of them.
Alaska-Aleutian megathrust
The Aleutian Arc
The Pacific Plate (actually, it’s a small portion of an ancient plate called the Kula Plate—much of which is deep beneath Alaska) is being forced beneath Alaska in an exceptionally long 3,600 kilometer arc. It’s well known that this interface has caused massive earthquakes. The tsunami from these quakes then range across the entire Pacific, striking Japan, New Zealand, Hawaii, and the California coast.
Much of the Alaska-Aleutian megathrust from its far end at the International Date Line to a point near Anchorage has ruptured through the 20th Century in a series of Mw7 through Mw9 quakes, unlike Cascadia. Its most infamous quake was the 1964 Alaska Quake, striking on Good Friday of that year. Tsunami rolled across the Pacific, killing a dozen people in California. The quake, an M9.2, is the largest quake in North American recent history. Long subduction zones tend to be segmented. Different parts of the zone move at different rates than others, and are stuck at different levels than others. This leads to areas that rupture and areas that are persistently stuck. We call these seismic gaps.
The hypothesis is that an area of an active fault system classified as a seismic gap hasn’t ruptured in a long time, and may do so in the near future (near future in geological terms could be tomorrow or centuries from that point in time.) The hypothesis has some considerable criticisms, but “seismic gaps” have been filled since the research was published. Such a gap exists along Alaska, at a place called the Shumagin Islands, in the eastern Aleutian Islands. Research indicates large quakes occur here with some regularity every 60 to 100 years or so, therefore we’re within the window, as it were. The last large quake here was in 1938.
Cascadia
Cascadia has become quite famous as of late, and the story of how its hazard came to be known is one of the neatest detective stories in science in quite some time, at least I think so. We’re going to take a good look at it, and
we’ll need our time machine again. All aboard, and everyone, remember; please keep all fingers and toes inside the device.
We’re going back to January 26, 1700, 9pm local time, to a place that later will become Newport, OR.
We won’t land---we’re not to disturb the people in the coastal village below. And they can’t see us anyway as we’re cloaked.
We can see below that the ground is shaking. Switching our display to night-vision, we can see all along the shore the land bubbling up with water. The people in the village are running out of their homes, grabbing their children, and heading for higher ground even as the ground continues to rock. The shaking goes on—we time it at almost six minutes. It takes that long for the margin, an immense area from northern California at Cape Mendocino to a point well past the future location of Vancouver, British Columbia to move. It’s massive---rocks are sliding off of the hills. A massive rock slide may have dammed the Columbia River. The marsh has subsided—it’s flooding with icy cold ocean water.
It’s not enough time for everyone to move to safety, down below. Sorry folks, we can’t help them. There shall be no temporal paradoxes. All we can do is watch in horror.
The sea has risen, as a massive earthquake has displaced the seafloor. It slams into the coast mere minutes after the earth has ceased shaking. It penetrates deep inland up bays and estuaries. Those who couldn’t run fast enough drown, and more will die from exposure. The survivors will tell the story of this evening and pass it down to their descendants, providing modern folks with an important clue. Sometimes myths are not always myths.
Another important clue is now racing across the Pacific—the tsunami---, where it will slam into Japan 9.5 hours later and surprise many chroniclers who felt no earthquake, and were experiencing no storm. It’s an orphan tsunami, and that too is an important clue. In the 1990s, it will allow us to pin down our earthquake’s time and magnitude.
And a third clue is left. The tsunami has left a distinct sand and mud layer where it overran the shore. It even snuffs out and buries firepits near the villages it’s washed away. At the same time, whole coastal forests have sunk below the tide-level as the earth subsided during the quake, and this is an even more important clue.
Did you know our time machine is also a submarine? It is! We’re headed down to the seabed to see the last clue. We call these clues turbidities. Basically, they’re landslides underwater, and we can see huge ones as we “swim” in our time machine-submarine up and down the Cascadia margin. These are new, but they bury dozens of older ones.
Back to the future! Hands and feet inside the time machine!
It’s the 80s now and people are puzzled. There’s a gap in the Pacific Ring of Fire. It’s Cascadia.
When you look at plots of moderate to large earthquakes, one will notice that there’s a ton in California. There’s a ton in Alaska, and westward through Kamchatka, Japan, Taiwan, the Philippines and so on. There are even quite a few in the intermountain region of the US. Not so much in Cascadia. There are more in Washington State, including some large deep ones in Puget Sound, but not exceptionally many. Oregon appears almost aseismic. Subduction was assumed in the 1980s despite an apparent lack of a trench (the Fraser’s and the Columbia’s immense outputs have buried it in “junk”), but it was assumed to be dead, or “creeping aseismically.”
But our clues are coming to light. In Japan, groups are reassessing the nation’s tsunami history and come across the Orphan Tsunami of late January, 1700. In Oregon and Washington, the sunken forests are noted, and carbon dating places the time of their sinking to roughly 1700 (tree ring studies indicate they stopped growing in 1699). The landslides on the floor of the ocean are noted. The tales from the Natives are taken into account. Further study is showing that the apparent quiescence really isn’t so.
Other clues come to light. Roads have moved slightly in Washington and Oregon. The mountains on both sides of Puget Sound are found to be moving closer together. The Coastal ranges are rising and being ever-so-subtly tilted toward the east. The region is being deformed.
The “junk” outputted by the Columbia River gave yet another clue. 7,000 years ago a volcano in Oregon exploded. Its explosion was so forceful it obliterated much of the mountain, ash and debris falling as far east as Saskatchewan. The ash layer is an important geological marker.
Rain then washed the ash layer that blanketed the region into rivers. The rivers met up with the region’s major rivers and dumped the ash into the sea, providing a very important layer. We know when Mt. Mazama erupted and we know the composition of its ash is very unique. Therefore, when the ash layer gets buried, we can tell when that mud was deposited.
Marine scientists found forty one separate landslides above and below the Mazama ash layer. That’s evidence of 41 Great Earthquakes, stretching back some 10,000 years. They didn’t find this at just one location. They found it up and down the coast. 19 of the 41 ruptured the entire plate interface from California to the British Columbia/Alaska border. Cascadia is merely asleep, not dead. And the strain is rising.
from Oregon.gov
One thing one can note is the great earthquakes come in clusters. It is not known if Cascadia has entered a period of prolonged quiet or not. It has been 313 years however since a large quake of any size has ruptured the fault, so we shall see. We’re approaching the “window” as it were, and the chance of a major quake is about one-and-three for any given year.
Oregon and Washington are being proactive, however. Coastal towns are constructing elevated shelters, tsunami drills are frequent, and building codes are being strengthened. The region’s apparent aseismicity however means there’s a long way to go before they’re fully ready, as soberly and eloquently described in an Oregon report published this year. If and when the quake happens, there will be little time for a tsunami warning to be disseminated for the people living along the coast. The warning will be the earthquake itself, and this is an important educational note.
Lastly there’s an interesting and intriguing side note. Large earthquakes ARE occurring beneath Cascadia in southern British Columbia and Washington state. But they take months to complete. They are also regular, almost predictable occurrences, with a timescale of 14 months. This is called“Slow Slip.”It occurs below the locked portion of the plate interface. It’s very interesting. And mysterious. Are slow slip events harbingers of the future? No one yet knows.
Puerto Rico and the Lesser Antilles
People’s ideas of paradise differ, but mine involves palm trees, crystal clear waters, hot temperatures, and sandy beaches. Like those of Puerto Rico and the other islands of the Caribbean.
The Caribbean also sits on its own tectonic plate and its collision with the North and South American plates has created significant and complex tectonics. The three plates (and along its margin, the bumping and grinding has fractured both plates into a series of independently moving microplates) are converging as well as sliding past each other very slowly. The crust here is very old—100 million years, and the rate of convergence is very slow.
From northern Hispaniola to the Virgin Islands, and then extending southeastward, is a subduction zone. Off Puerto Rico is a well-known trench—the deepest in the Atlantic. The subduction zone has a well-developed island arc—the islands of the Caribbean. The trench the subduction of the Atlantic crust has formed is buried by the “junk” outputted from the Orinoco River in South America. We call this an accretionary wedge. The sediment has also obscured the location where the Caribbean, North American, and South American plates meet (this is called a Triple Junction.) We have no idea where it really is, and I’ve seen it suggested that it doesn’t exist and North and South America simply sit on one gigantic ,westward moving plate.
The subduction zone was given new attention after 2004 because of its similarities with the zone off Sumatra, and earthquake hazard across the Caribbean is now big news after the January 12, 2010 Haiti quake. That quake did create several tsunami-one on the southern shore of Hispaniola was generated by underwater landslides, and another in the Gulf of Gonaives generated by uplift of the seabed adjacent to the coast.
Elsewhere, an intense earthquake sequence north and east of the Dominican Republic that began in 1918 and ended in the early 1950s (some consider an M6.4 quake in 2003 near Puerto Plata to be part of this sequence) generated several tsunamis, several of which show up on tide gauges in the eastern United States. An 1867 earthquake (it was actually 2, in close succession) near St. Thomas caused a tsunami so damaging it delayed the American acquisition of the islands for fifty years. Earthquakes in 1690, 1839 and 1843 caused local tsunami from Antigua to Martinique. But large earthquakes are infrequent here, and forgotten, leaving tens of millions somewhat unaware of the hazard beneath their feet and beneath their seas.
Let’s take a look at this map
Some of the big megathrust earthquakes around the Caribbean.
I’ve plotted (from research) the areas of the subduction zone that have ruptured. The February 8, 1843 earthquake between Guadeloupe and Antigua was once thought to be below M8—
new historical data from archival newspaper records indicates it was so big it rattled buildings and nerves from Guyana to Bermuda to Washington D.C. to New York City. This is a quake that likely approached M9. It didn’t, apparently, rupture the seabed, thus no tsunami rolled across the Atlantic, although local deformation did cause waves in the islands and subsidence on Antigua and Guadeloupe. Just to the south, a massive quake near Martinique occurred in 1839, and probably was around M8. To the south between Martinique and Trinidad, there are no known large megathrust earthquakes. Either the area is locked and currently loading, like Cascadia, or it’s actually aseismic (although fairly large deep quakes occur, including one near St. Lucia in 1953). Further research using GPS data is needed (and I believe, underway).
The segment of the megathrust near Hispaniola’s north coast has ruptured, in a series of earthquakes starting in 1918. But north of Puerto Rico and the Virgin Islands, there are no known historical megathrust events (quakes in 1785 near the Virgin Islands and 1787 near Puerto Rico are now thought to be “outer rise” earthquakes.)
The presence of this trench and the massive Indian Ocean quake in 2004 is a source of concern for the US East Coast, so much so that a tsunami warning system has been created for the Atlantic, and tsunami wave heights from a hypothetical have been modeled.
Simulated tsunami waves from a hypothetical 8.6 Puerto Rico Trench Earthquake, from NOAA.
However, the Caribbean Plates and North American plates are sliding past each other, with only a small component of the North American plate diving beneath the islands. The top of the trench is riddled with landslides. It is not known whether this region can produce an M9, and research continues. Some early research indicates that just before Columbus arrived or perhaps in 1755, a
massive tsunami may have rolled over some of the Virgin Islands.
Like Cascadia, the warning in Puerto Rico and the Virgin Islands will be the earthquake itself.
Landslides
I mentioned landslides underwater before. They too can cause tsunami.
Up and down the US East Coast, just at the location where the continental shelf drops off into deep ocean, are old scars. These scars are ancient landslides, and above those scars are unstable sediments that could perhaps, one day, give way.
The rivers of the East Coast dump a lot of junk into the ocean. In addition to our trash and poop both human and agricultural, they dump considerable amounts of sediment onto the continental margin. Along the US East Coast this margin is considered “passive.” There’s no plate boundary here. It’s supposed to be nice and quiet. It sometimes isn’t.
In 1929 a M7.2 earthquake struck beneath the seabed south of Newfoundland. The earthquake was widely felt across New England and the northeast, and much of eastern Canada. The quake dislodged a massive amount of sediment that cascaded down the continental shelf into deep water, severing several telephone cables along the way and generating a tsunami that slammed into the Burin Peninsula. 29 people were killed. The tsunami did roll across the Atlantic, registering itself on tide gauges in Atlantic City and in Portugal, but the waves were harmless by that point.
The presence of these landslides has caused the USGS and the Nuclear Regulatory Commission to commission some serious research into tsunami hazards on the US East and Gulf Coasts, which you can read in its entirety here.
Many of these landslides appear to have occurred prior to 7,000 years ago, or around the time of the end of last ice age.
Landslides are also noted around the floor of the Gulf of Mexico. Research continues here as well.
Things falling into the sea can displace the ocean. A truly massive slide of rock slid into Lituya Bay in Alaska following a significant earthquake in 1958. The wave it generated was 1,000 feet high in the narrow bay. Luckily, it dissipated in the open ocean, but it scoured the hillsides clean. Other ancient slides off of Norway flooded portions of the British Isles thousands of years ago.
The Canary Islands
The Canary Islands are a volcanic hotspot off of Africa. The islands also have, periodically, collapsed into the Atlantic. La Palma is one such island where it’s believed one day, a significant portion of the island’s volcano will fall into the sea, generating truly massive megatsunami.
Or will it?
Since Ward and Day’s original research was published (it had a wave 25 meters high approaching Florida), subsequent research has taken that hypothetical wave down several pegs. Because of the type of water displacement (from above rather than from below the seabed), any wave generated would have its power slowly dissipated as it rolled across the ocean. It’d be devastating to the Canaries and nearby African coast, but it’d be less so in the Caribbean and probably not damaging at all when it reached the United States.
With a near-field landslide, if it’s caused by an earthquake, the earthquake will be the warning itself. The wave however, will move fairly slow through the shallow waters of the continental shelf, giving perhaps two to three hours before the waves strike the coast. It’d be akin to a Category 3 to 4 storm surge.
Asteroids/Meteorites
We’re back in our time machine and we’re going to a place we’ve visited before: the New York City area. Last time, we went back to 1315.
This time, we’re going back to 300BCE.
We’re hovering above the tip of Manhattan Island. There are swamps and tidal flats. We watch a young man staring out into New York Bay.
For no reason the sea has begun to recede. Not long before, perhaps, there was a flash on the distant horizon and perhaps a boom.
The sea returns in force and it’s too late for him to run. A wall of water far higher than any storm surge has already overtopped the barrier islands of Long Island and New Jersey and is now forcing its way up the Hudson. It’ll travel inland up the Hudson estuary as far as the future site of Albany.
And it’ll leave behind two things: A tell-tale sign that paleotsunami hunters look for—a sediment layer farther inland than any storm surge, and perhaps tiny minute little diamonds, which meteorite impact hunters look for. But this is in considerable debate. The tsunami absolutely did happen—its sediment layer is clearly marked in multiple locations, but the tiny minute diamonds, called tektites, are in dispute. The same tsunami could have been caused by a landslide down the continental shelf off the shores of Long Island.
The Holocene Impact Working Group spends its time looking for evidence of ancient (but not that ancient) impacts. It’s an interesting group, with interesting geologists and others and is sort of at the fringes of science at the moment. But with most of Earth covered in water, the likelihood of a meteorite impact in the ocean is pretty good, if it happens. And if it happens, it will generate massive waves that will devastate any coast they come in contact with.
Let’s hope we never see it.
Conclusions
I hope this has been a good backgrounder on tsunami and their causes, and the areas that directly threaten the United States.
If you’re on the beach and feel a big earthquake, leave the beach as soon as possible for high ground.
Here is an example of a test tsunami warning:
WCATWC Message #2
WEXX20 PAAQ 021335
TSUAT1
BULLETIN
TEST...TSUNAMI MESSAGE NUMBER 2...TEST
NWS WEST COAST/ALASKA TSUNAMI WARNING CENTER PALMER AK
935 AM AST THU APR 2 2009
...A TEST TSUNAMI WARNING IS IN EFFECT FOR PUERTO RICO AND THE VIRGIN ISLANDS...
...THIS MESSAGE IS INFORMATION ONLY FOR U.S. AND CANADIAN ATLANTIC AND GULF OF MEXICO COASTAL REGIONS NOT INCLUDED IN THE AREAS LISTED ABOVE...
RECOMMENDED ACTIONS
A TSUNAMI HAS BEEN GENERATED WHICH COULD CAUSE DAMAGE TO THE
WARNING AND/OR ADVISORY REGIONS LISTED IN THE HEADLINE. PERSONS
IN LOW-LYING COASTAL AREAS SHOULD BE ALERT TO INSTRUCTIONS FROM
THEIR LOCAL EMERGENCY OFFICIALS. EVACUATIONS ARE ONLY ORDERED BY
EMERGENCY RESPONSE AGENCIES.
- PERSONS IN TSUNAMI WARNING COASTAL AREAS SHOULD MOVE INLAND TO
HIGHER GROUND.
THIS MESSAGE IS BASED ON EARTHQUAKE DATA... OBSERVED TSUNAMI AMPLITUDES... HISTORICAL INFORMATION AND FORECAST MODELS.
A TSUNAMI HAS BEEN OBSERVED AT THE FOLLOWING SITES
LOCATION LAT LON TIME AMPL
------------------------ ----- ------ ------- -----------
CHARLOTTE AMALIE US VI 18.3N 64.9W 1315UTC 0.77M/2.5FT
CULEBRA ISAND PR 18.3N 65.3W 1317UTC 0.93M/3.0FT
LAMESHUR BAY ST JOHNS 18.3N 64.7W 1322UTC 1.05M/3.5FT
TIME - TIME OF MEASUREMENT
AMPL - TSUNAMI AMPLITUDES ARE MEASURED RELATIVE TO NORMAL SEA LEVEL.
IT IS ...NOT... CREST-TO-TROUGH WAVE HEIGHT.
VALUES ARE GIVEN IN BOTH METERS(M) AND FEET(FT).
PRELIMINARY EARTHQUAKE PARAMETERS MAGNITUDE - 8.6
TIME - 0900 EDT APR 02 2009
0900 AST APR 02 2009
1300 UTC APR 02 2009
LOCATION - 19.0 NORTH 65.0 WEST
65 MILES/105 KM NE OF FAJARDO PUERTO RICO
80 MILES/129 KM NE OF SAN JUAN PUERTO RICO
DEPTH - 12 MILES/20 KM
THE PACIFIC TSUNAMI WARNING CENTER IN EWA BEACH HAWAII WILL ISSUE MESSAGES FOR AREAS IN THE CARIBBEAN OUTSIDE PUERTO RICO AND THE VIRGIN ISLANDS.
TSUNAMI WARNINGS MEAN THAT A TSUNAMI WITH SIGNIFICANT WIDESPREAD INUNDATION IS IMMINENT OR EXPECTED. WARNINGS INDICATE THAT WIDESPREAD DANGEROUS COASTAL FLOODING ACCOMPANIED BY POWERFUL CURRENTS IS POSSIBLE AND MAY CONTINUE FOR SEVERAL HOURS AFTER THE INITIAL WAVE ARRIVAL.
THIS MESSAGE WILL BE UPDATED IN 30 MINUTES OR SOONER IF
THE SITUATION WARRANTS. THE TSUNAMI MESSAGE WILL REMAIN IN EFFECT
UNTIL FURTHER NOTICE. REFER TO THE INTERNET SITE
WCATWC.ARH.NOAA.GOV FOR MORE INFORMATION.
AMZ712-715-725-735-742-745-PRZ001>003-005-007
Tsunami warnings in the US come from the
Pacific Tsunami Warning Center and the
West Coast and Alaska Tsunami Warning Center . The WCATWC is also responsible for Puerto Rico, the Virgin Islands, the US and Canadian East Coasts and the Gulf Coast. The PTWC is responsible for the Indian Ocean, South China Sea and Caribbean Sea outside of the United States’ claimed waters.
Thanks for reading.
(also, this diary directly inspired the writing of this one. Thanks, OceanDiver!)