Where there are volcanoes, there are earthquakes. Both are a sign of a broken earth. Volcanoes require vertical movement and earthquakes (by and large) are horizontal: the two are not identical, but to get a volcano you need a vertical path, and to get that you need to move crust sideways. Enter the earthquakes. A spreading centre can just push the crust aside, with minimal earthquake activity, but still not entirely silent. In contrast, a subduction zone really goes for it. Here, of course, the earthquakes and volcanoes do not coincide. The earthquakes are close to the subduction trench and the volcanoes are further down, where the subducting plate reaches the depth where melt can occur, with the molten rock percolating up. The volcanoes may appear 100 km from the subduction region. And the subducting plate pushes up the continental plate it subducts under, forming its own mountain ranges. Sometimes the volcanoes grow on top of those ranges, sometimes they are distinct. And the weight of the mountains can fracture the earth near them. A jumble of faults may form. Indeed, volcanoes are like acne: they are symptoms of the turmoil beneath. The earth too suffers its puberty.
It is not a surprise that people in volcanic regions need to be earthquake-aware. And as subduction zones tend to be coastal, there is an added danger of tsunami activity. Furthermore, these are seldom flat deltas -an active earth builds relief, and relief with shaking gives the risk of landslides. Hazards seldom come alone.
Alaska has it all. There is the wonderful ocean, teeming with life in an abundance (though not colour) that puts the tropics to shame. There are the mountains and the wilderness, beautiful to see but harsh and unforgiving. There is the eternal winter of Denali. There are volcanoes to admire (from a safe distance). It is tourist heaven. But to live there requires coping with everything the earth can throw at you, from the certainty of winter to the suddenness of an earthquake. Living here requires -and builds- character. And now, Alaska is back in the news: after months of rumblings across the state, and a larger earthquake in the normally earthquake-free zone in the north, a big one hit the most populated area of the state. The rumblings had indeed given plenty of warning that something could be brewing. And still it came as an early-morning surprise.
Alaska is a late addition to North America. That is true politically, with the purchase from Russia in 1867. At the time Russia was recovering from the Crimean war and it was expected that in any resumption of hostilities Britain would capture Alaska. That prospect alarmed both Russia and the USA, so the purchase suited them both. How little times have changed, with another Crimean war and a North Atlantic trade war in progress! The geology of Alaska also derived from Asia. The land was build up north to south. 400 million years ago, the sea between the Pacific and the Arctic ocean was shallow, an extension of an older mountain range reduced to small islands and submerged land. Over the next two hundred million years, a tectonic plate drifted north and hit the shallows. This collision formed the Brooks mountain range 150 million years ago, and it largely closed off the connection between the Pacific and the Arctic ocean. Later, a new plate fragment hit, and the White Mountains formed in the same way, separated from Brooks by the Yukon valley. Finally, the Alaskan range formed when the current Pacific plate began to subduct underneath the young continent.
Of the faults the riddle Alaska, best known is the subduction trench. The Pacific plate is moving north at about 5 cm per year. As it sinks, it forms the Aleutian trench which runs from Alaska to Russia.
The image shows how the process works. The Pacific plate comes in from the right (south). As it arrives, a cover of accreted rock is scraped off and left behind, pushed up against the continent. This is fairly weak rock with many fault lines. These are called ‘splay faults’ because the stress of a single fault system is spread over many individual faults. Montague island, off the coast of Alaska, formed this way. Further along, an area of subsidence forms. Finally, the subducted plate meets the much stronger rock of the true continent and slides underneath.
The zone between the two plates, where the continent slides over the subducting oceanic floor, is where the major earthquakes occur. The events are often close to the trench, but along the Alaskan coast they also occur further north: the angle of subduction is not as steep here as it is further west along the Aleutian arc. Note how the trench peters out south of the Cook Inlet, also an indication of the shallower subduction. In Alaska, the subduction is less about sinking ocean floor and more being overridden by the lower-density rock of Alaska. The line of Alaskan volcanoes ends at Cook Inlet, and doesn’t resume until near Canada, with the enigmatic Wrangell volcanic field. This is called the Denali volcanic gap. Everything is connected: the missing trench, the shallow subduction and the missing volcanoes. What is missing is the cause. The suggestion is that the oceanic crust here comes from an oceanic plateau, which formed in the Pacific and drifted north where it met Alaska. The plateau is much thicker than usual oceanic crust, at 20 kilometers. This thickened slab is called the Yakutat crust, and it currently lies underneath the area west and north of Anchorage.
The oceanic slab descend at an angle of 30 degrees until it reaches 100 km in depth, after which it goes down much steeper to 400 kilometers. The Alaskan Range, the mountain range that includes Denali, forms an arc where this subducted slab descends steeply; the mountains are pushed up by the buoyant oceanic plateau. Alaska is a well squeezed place.
There is a second fault that is important. The Denali fault follows the arc of the Alaskan Range. Movement along this fault is limited to about 1 cm per year westward. But it builds up a lot of stress. On November 3, 2002, one of the the largest on-land earthquakes known in North America happened here, when 220 kilometers of the fault ruptured in an M7.9 event. The Denali fault has a large bend in it, and Denali itself lies just south of this bend: perhaps the bunching of crust along this bend has pushed Denali to its extreme altitude.
In between the subduction trench and the Denali fault, a part of Alaska has become semi-separated from the North American plate. It forms the Wrangell sub-plate. Anchorage is in fact located on this sub-plate, connected to North America but not fully part of it. It seems fitting.
There are more faults here. In between Denali and the trench is the 200-km long Castle Mountain fault, which runs north of the Cook Inlet: part of it is visible as a scarp on the surface. In 2007, a report came out showing evidence that earthquakes occur on this fault abut every 700 years, with the last one 650 years ago. The authors predicted that an event of M6 to M7 would be possible on this fault in the next century. There are other faults in the area, such as the Border ranges fault which runs east of Anchorage.
The region shows a complex structure, with a major fault through the mountains, shallow faults running parallel to it, and a subducting plate underneath. Each one of the faults and boundaries is subject to earthquakes. Is it a wonder that shaking here is frequent?
A shaken history
A map of major earthquakes around southern Alaska shows that most of them occur towards the Aleutian arc, and are in between the coastal mountains and the subduction trench. The trend picks up again along the North American continent, were the major events closely cluster around the main fault. In between, things are much more complex. Here, some events occured within the Gulf of Alaska, showing the stress in the oceanic plateau. Two events were near the Denali fault, and four were in-land along the Cook Inlet and the Prince William Sound. Things are far less predictable here.
The map above shows the largest earthquakes in the Cook Inlet area over the past century. Two are on the Denali fault (1912 and 2002); the others are distributed around the area. There is no obvious connection to the known faults in the area: either the earthquakes come from faults that are not visible on the surface, or (more likely for large events) they occur on the frictioned surface between the continental crust and the subducting plate.
Plot the M5.5+ earthquakes, and a bit more of a pattern appears. Some follow the Denali fault. Others are the in middle of the Cook Inlet. There is a cluster on the west side of the Cook Inlet, and a few follow the Castle Mountain fault.
The depths are also clustered. The Denali earthquakes are around 5-20 kilometers deep although two are at 130 kilometers. The ones near Anchorage are around 30-60 kilometers and at the top of the Inlet they are mostly around 100 kilometers deep. There are few strong earthquakes in the upper crust: most earthquakes trace the descending slab. The subducting plate lies about 40 kilometer below Anchorage where it is much shallower than around the top of the Cook Inlet. The earthquakes caused by the subducting plate are thrust quakes.
The land region around Anchorage, including Denali, has around one M6.5+ earthquake per decade. All these quakes would have been major disasters anywhere else in the world. But Anchorage is build to cope. It has experience. It may be shaken – but not stirred.
The great quake of 1964
But one of the earthquakes was too much even for Anchorage. The stress had build up for a long time. The process was the same as the thrust earthquake described above. The subducting Pacific plate had gotten caught by the plate above. Friction kept the two locked underneath Anchorage. In this situation, the subducting plate keeps moving: it has little choice, with the whole mass of the Pacific behind it. So the continental crust, locked in place, was pulled down with it. The effect was that the south side of Anchorage was being pulled down and the northern side pushed up. This had happened over many years, so people were not aware of the slowly changing landscape. Until, on March 27, 1964, at 5:36pm local time, 25-kilometer below the surface, the connection between the two plates gave way. Over the next 4.5 minute, Alaska snapped back into place. The result was the largest earthquake ever measured in the US, and the second largest known worldwide. The magnitude of the event was a staggering 9.2.
Thrust earthquakes of this size are called mega-thrusts. There is no strict definition, but a usable one is to call a thrust quake of M8+ a mega-thrust. What causes a thrust quake to reach this size? That is not fully known. The speed of the subducting plate is obviously a factor. So is the subduction angle: shallower subduction gives rise to stronger thrust quakes. In fact, this mechanism is the only one that can produce earthquakes this size, so all M8+ earthquakes are mega-thrusts. Five earthquakes of M9+ have been recorded in history: Kamchatka 1952, Chile 1960, Alaska 1964, Sumatra 2004 and Japan in 2011. Casualties ranged from an estimated 5,000 for the Chilean earthquake to 250,000 for Sumatra. In this list, the Alaskan quake is exceptional for the very low number of people who lost their lives, 131. Many of the casualties were in the coastal village of Chenega which lost a quarter of its population from the tsunami. At the airport, a controller died when the tower collapsed. But in view of the extreme size, the remarkably low number shows how well prepared Alaska is.
Still, the effects were severe. In four minutes, a century of deformation was undone. The area to the south, which had been pulled down, was uplifted: the highest uplift was close to 15 meter. The subsidence along the sound was less, but still over 2 meters in places. A tsunami formed with a highest run-up height of 65 meters. The wave reached as far as Hawaii. The worst local effects were from the multiple landslides.
The Anchorage earthquake occurred in the early morning of 30 November 2018. In hindsight, it began on Jan 5, 2018 when an M7.5 hit off the coast of Canada. This is in itself not that unusual. However, it was followed by an M7.9 earthquake off the Alaskan coast, which occurred on 23 Jan. Were the two events related? They seem far apart for that, but such a quick succession of major earthquakes is unusual, even here. One may have triggered the other if it was already primed to go.
And on Nov 30, the earth just north of Anchorage shook when the locking of the continental and subducting plate failed. It was recorded at M7.0, minor compared to 1964 but massive on any other scale. A link between the January and November events seems speculative. However, there were months of minor earthquakes in the general area since the summer of 2018. It seems plausible that the January event transferred some stress, and that the area was already prone to slipping. The minor quakes were a danger sign, warning that a larger quake was coming. Foreshock events here are not uncommon. The M7.9 earthquake in 2002 was preceded by an M6.7 earthquake which occurred 10 days earlier.
Initial reports suggests that the damage from the 30 November earthquake is remarkably limited. Some damage to some buildings was mentioned, and some roads, specifically Glenn Highway, Vine Road and Minnesota Drive, are badly cracked, possibly affected by liquefaction. The damage is estimated as over 100 million dollar. For such an event, that seems very low. But no matter how well prepared the area is, for people in it is a terrifying experience, made worse by the aftershocks.
The 2018 earthquake happened at almost the same location as the great 1943 earthquake, which coincidentally also was in November but was much stronger at M7.6. The depth of 40 kilometer makes it a thrust quake, with the same process as the 1964 mega-thrust but much smaller. As I write this, the aftershocks are still continuing, at depths of 20-40 kilometers. The aftershocks are all north of the main event, between the Inlet and the Castle Mountain fault. But it hasn’t triggered any movement on the latter fault which is a good thing given the possibility that it may be close to its own event. In any case, typically the largest aftershock is one magnitude less than the main one. The largest aftershock so far is less than that, and so it is possible that there will be another M6 quake in the next few days.
Normally, a quake this size resolves the stress for quite some time. The repeat time can be hundreds of years or longer. But not here where M7-size quakes seem a dime a dozen. And some earthquakes of this size in Anchorage come in pairs, around 10 years apart. This happened in 1899 and 1909, (a further event in 1912 was near the Denali fault), and in 1934 and 1943. After the 1964 mega-thrust, it was almost 20 years before there was another M6.5 thrust, in 1983. And now, 35 years later, we have had the next thrust. (A few earthquakes of similar size in between were either much deeper or much shallower, and of different type). Before 1964, there had been 11 thrust earthquakes M6.5 or larger over 50 years, or two per decade. After 1964, there were 2 over 50 years. It shows how quiet things became.
So what is next? Have the quiet times after 1964 come to an end? If so, we can expect such thrust quakes to become more frequent. The next one will not be 35 years away but may well be within 1 or 2 decades. But whatever happens, Alaska is remarkably well prepared. It will manage. It knows that the beauty of this land of sea, mountains and volcanoes is build on broken ground. It is part of the package.
Albert, December 2018