When two earthquakes with magnitudes of 7.6 struck the Latin American country of Peru on Tuesday evening (local time) within five minutes of each other, one had to fear for the worst. After all, in contrast to its southern neighbor Chile, Peru is not known for its earthquake resistant structures and strictly enforced building codes. But, although the shaking of both quakes was widely felt across western South America, it caused hardly any damage. Even in the epicentral region in the sparsely populated Peruvian jungle near the Brazilian border to the east of the Andes, no buildings collapsed. The head of Peru's emergency services, Alfredo Murgueytio, was quoted by Reuters as saying that "there are no damages reported." Several residents of the Brazilian city of Brasileia, 150 miles east of the epicenter, told the same news agency, that they felt the ground shake and that chairs and tables rattled during the quake, but that there was no visible damage.
At first glance, the absence of any significant damage was very surprising. After all, the single quake which occured earlier this year in Nepal wreaked havoc in Kathmandu and other areas of the Himalayas . It had a magnitude of 7.8 and hence was only slightly stronger than Tuesday's double temblor.
What is so different about these two earthquakes in Peru and the one in Nepal? The Nepal quake caused massive destruction with over 9000 casualties while the doublet on Tuesday hardly damaged anything at all. The main reason for the mild effects of the two Peruvian earthquakes was their focal depth. While the Nepal earthquake had its origin less than 10 miles below the surface, seismologists at several seismological centers around the world calculated the depth of the Peru double quakes to be between 375 and 400 miles below the surface. By the time seismic waves from such deep earthquakes reach the surface, they have lost a lot of their initial energy.
Think about what would happen when an earthquake with a magnitude of 7.6 were to occur in the Los Angeles area. While we are sure that it will cause quite a bit of damage in the epicentral region down south, we do not expect this earthquake to cause significant structural damage in the Bay Area. The seismic waves may be felt in our area around the San Francisco Bay, but they will have lost much of their punch on their almost 400 mile long journey way northward. It so happens that the distance between San Francisco and LA is similar to the depths of Tuesday's earthquakes in Peru.
While deep earthquakes like the two on Tuesday under Peru are usually rather gentle on buildings and structures on the Earth's surface right above their focus, they can be felt over a surprisingly large area. A magnitude 8.2 quake, which had its source almost 450 miles below Bolivia in 1994, made skycrapers in Toronto sway. Similarly, the deepest earthquake ever measured, the magnitude 7.8 quake in the Izu-Bonin-Trench in late May this year, caused a slight swinging movement of a 50 story building in Pasadena. The distances between Bolivia and Toronto and the Bonin Trench and Southern California are 4500 and 5500 miles respectively.
What is the reason for such ultra-long distance effects? In contrast to shallow focus quakes, the seismic waves of deep temblors do not have to travel through the Earth's crust on their way to distant locations. They start their journey in the Earth's mantle. Compared to the Earth's crust, this section of the Earth's interior is much more homogeneous, and therefore attenuates seismic waves much less strongly than the more heterogeneous rocks of the crust.
In summary: deep earthquakes usually cause less damage than shallower ones, but they can be felt over much, much longer distances. If you happen to be in a skyscraper which gently rocks from side to side, the motion is not necessarily due to the wind. The swaying of a tall structure can be caused by seismic waves from an earthquake continents away. (hra112)
|More than 400 small earthquakes have occured in the seismic swarm under San Ramon during the last two weeks (Map: USGS)|
There was an era in earth science more than thirty years ago, when earthquake prediction was at the top of the seismological research scale. After some apparent successes in China and other places, seismologists thought they were close to being able to tell the world exactly where and when devastating temblors would occur. Parkfield, a tiny hamlet in central California northeast of Paso Robles, was the epicenter of this research. There, where the San Andreas Fault is so strikingly exposed, dozens of scientists from the US and abroad placed hundreds of sensors into the ground, hoping to catch the latest precursor. Unfortunately, after more than a decade of enormous effort in Parkfield, earthquake prediction proved as elusive as the search for the Midas' touch to turn everything into gold.
One of the questions seismologists tried to answer was, if earthquake swarms could be the prelude to a Big One. This hypothesis got some notoriety in the summer of 1975. Six years before, the construction of the Oroville Dam north of Sacramento had been finished and this tallest earthen dam in the US had impounded the Feather River without any problems. But when suddenly in June 1975 earthquakes started to happen under the reservoir, authorities became worried. Would the quakes shake to dam to pieces? However, within a month the swarm had subsided and people breathed a little easier - until August 1, when a quake of magnitude 4.7 shook the dam. Bruce Bolt, then the director of the Berkeley Seismological Laboratory, was asked if there was more to come. He said it was possible and indeed, a few hours later, a magnitude 5.7 quake happened in the area of the swarm.
Not only scientists began to ask whether Bolt's statement was a true prediction or if the occurence of the strong quake under the dam was an untimely coincidence. Even though we seismologists have clearly failed to find any reliable precursor so far, the same question is being asked again today, as it was after the Oroville incident: Is the current earthquake swarm under San Ramon and Danville in the East Bay a prelude to something bigger? Starting with a 0.8 magnitude microquake on October 13, more than 400 temblors have since occured along the Calaveras Fault under the Crow Canyon Country Club. At least eight of these quakes had magnitudes greater than 3 and were clearly felt all the way to Concord in the North and San Jose to the South. As of this writing, the swarm goes on and on.
To say it very clearly: Earthquake swarms are not at all an indicator that something bigger is about to come. Too many times such swarms have happened without culminating in a big, destructive quake. This is especially true for the I-680 corridor. In May a swarm under Concord came and went. And since 1970 the northern segment of the Calaveras Fault in the region of Alamo, Danville and San Ramon has seen at least four swarms, most recently in 2003. Each lasted for weeks or months, but so far none was closely followed by a significant quake.
Does that mean we can exclude that a bigger, destructive quake will happen soon on this section of the Calaveras Fault? No, we can't! This segment of the fault has not had a significant quake since it last ruptured in the 1860's. Given the stresses continuously exerted on it by the movements of the tectonic plates here in the Bay Area, the likelihood of a quake of magnitude 6.7 or greater under San Ramon and Alamo in the next 30 years is about eight percent. At first glance this number seems very low - but it is not zero. Such a quake can happen today, next week, in thirty years or even later than that. And this is independent of whether the area is experiencing an earthquake swarm or not. (hra111)
Earthquakes don't happen accidentally. Instead, temblors occur only when a normally delicate balance of forces within the Earth's crust suddenly goes off-kilter. At least one of these forces, the tectonic push which is ultimately driven by the endless drift of the lithospheric plates, tries to break apart the rocks underneath our feet. Others, like friction, shear strength, or the weight of the overlying rocks, act as defensive linemen, counteracting the tectonic push. Most of the time, these and other forces are in an equilibrium. However hair-raisingly tight this equilibrium may be, its keeps the rocks to stay put and let's Mother Earth not produce any major quakes. However, the moment the plate tectonic drive becomes too strong, the defensive forces collapse and an earthquake occurs. The process leading to such a temblor is the same, no matter if the result is a small shaker along the Hayward Fault or a megaquake like the one off the coast of Japanese Honshu in March 2011.
If we seismologists could only measure in detail all of these forces or the various mechanical stresses each of them exerts on the rock, we would certainly be able to better forecast earthquakes. But such measurements are far from trivial. Take the Hayward Fault for example, which runs for almost fifty miles through one of the most densely populated regions in California. We would need to drill into the fault every few hundred feet or so. Each of these drill holes would have to reach to a depth of approximately eight miles and needs to be equipped with a string of a dozen sensors to measure the stresses. Given the current drilling and sensor technology, such a system could not be realized even if one throws all available money in the world at it.
Lacking the direct measurements we have to extract information about the stresses acting along an earthquake fault via various detours. However, none of these pathways has proven very reliable and the results are often contradictory. One of the strange paradoxes resulting from such stress investigations along faults is that even strong earthquakes don't necessarily relieve a lot of mechanical stress. Instead of resulting in a much less stressful environment, the actual stress drop caused by a temblor can be ridiculously low, even if when the quake had a large magnitude.
Jeanne Hardebeck, a research seismologist at the US Geological Survey in Menlo Park, has now looked at the stress regimes encountered in the areas, where the biggest earthquakes in the last two decades occurred on the planet. These are the subduction zones around the Pacific. While using the orientation of fault planes of smaller earthquakes which occur there, she finds that it actually doesn't take that much force to generate a megaquake. As she writes in the journal Science the frictional forces - the defensive linemen - holding together a subduction fault are actually surprisingly weak. However, in contrast to common intuition, the Earth's crust in such subduction zones is similarly weak. This means that the sum of the various forces acting on it are less strong than anticipated. Hardebeck's findings suggest that the Earth generates its most devastating earthquakes along rather weak faults which are located in very low stress environments. If these results get confirmed along other faults, it seems that we still have a long way to go to fully understand the interplay and the strength of forces necessary to generate an earthquake. (hra110)
The Hawaiian Archipelago, the quintessential island dream for many tourists and vacationers is also a unique "dream island" for Earth scientists. Its active volcanoes, most and foremost Kilauea, which has been continuously spewing lava for more than 30 years, are a unique window into the deeper regions of the Earth. There is, however, at least one scientific enigma about Hawaii and its fiery mountains: Why are there active volcanoes somewhere in the middle of the vast Pacific Ocean, thousands of miles away from the nearest volcanically active plate boundary?
|The plume under Hawaii is shown in yellow. It reaches from the boundary between the core and the mantle at the bottom of the picture (CMB) to a much shallower depth where it fans out. The green triangle on the top represents the Hawaiian volcanoes. (Source: 2015, Nature, 525, 95-99, doi:10.1038/nature14876)|
More than four decades ago, on March 5, 1971 to be exact, a geophysicist from Princeton, proposed a solution to this dilema. The volcanoes in Hawaii, wrote W. Jason Morgan in a paper in "Nature", are fed by a hot spot, a blob of hot magma coming from deep within the Earth. Since he published his hypothesis, dozens of features in Hawaii and at many places elsewhere on Earth were discovered, which strongly support the hot spot idea.
But again, one fundamental question was left unanswered: Where are the roots for these rising magma plumes, which feed the volcanoes. Even after fourty years, this is still the subject of a vigorous, sometimes even heated debate among Earth scientists: Do the plumes feed from magma reservoirs in the upper mantle of the Earth at a depth of a few hundred miles, or do they reach all the way to the boundary between the Earth's mantle and its core, 1800 miles beneath our feet. And also: What are the actual shapes of such plumes? Are they shaped straight like chimneys or do they fan out somewhere on their way to the surface of the Earth.
The answers to these questions, as academic and esoteric as they look at first glance, are of great importance in our still rather limited understanding of the heat engine which drives plate tectonics and hence generates deadly earthquakes and tsunamis. If the plumes have shallow roots, the mechanism pushing the plates all over the Earth's surface does not involve the whole mantle. A deep root on the other hand would indicate a huge heat engine with convection currents all the way down to the Earth's core.
|Supercomputer simulation of plumes of hot rock rising through the mantle to the surface, where they generate volcanic eruptions that form island chains. Animation by Scott French, NERSC & Berkeley Lab; video by Roxanne Makasdjian and Stephen McNally, UC Berkeley|
Using the recordings of almost 300 strong earthquakes, which shook the globe over the past 20 years, and weeks of computing time on a cluster of supercomputers at the Lawrence Berkeley Lab, two reseachers from our own Berkeley Seismological Laboratory seem to have found answers. In today's issue of "Nature" Barbara Romanowicz and Scott French describe the plumes in detail: They clearly originate at the core-mantle boundary. The researchers were surprised when their computations showed that the plumes are not as thin as a chimney but are at least 400 miles across in their lower part. At a depth of about 600 miles the plumes become even broader and begin to fan out. The actual hot spots at the surface - the volcanoes - are then fed from these spread out magma blobs.
The technique, which Romanowicz and French used to peer into the interior of the Earth in such detail, is called seismic tomography, which is similar to a CT-Scan in medical diagnostics. While the mathematical techniques to generate three dimensional pictures of something normally invisible are the same in imaging the innards of a human body and the inside of the Earth, there are differences: The X-rays used in medicine are absorbed differently by different tissues in the body. In the Earth's interior, however, seismic waves slow down slightly, when they pass through a hotter medium - like the magma inside a plume. These tiny differences in the travel times of all kinds of seismic waves are the basis for the current research. Buy using the supercomputers, the two researchers converted them into the first ever complete CT-Scans of the plumes under the Pacific Ocean and other volcanic regions on Earth. (hra109)
|The black arrows show how far the East coast of Honshu moved during the Tohoku earthquake in March 2011. Some sections (longest arrows) shifted more than ten yards (Source: Ozawa et al., 2011, Nature, doi:10.1038/nature10227)|
When on April 25th the Himalayas were shaken by a 7.8 magnitude quake, more than 8700 people died and the Nepalese capital Kathmandu and many villages in the mountainous country-side suffered catastrophic damage. The quake also triggered a huge avalanche on Mt. Everest, the world's highest mountain. It rushed through the climbers' base camp, killing at least 18 people and ending the annual climbing season, which had opened only a few weeks earlier. However, had the climbers been able to get up to the summit of the world's tallest peak, they would not have found it where it used to be. Scientists from the Chinese National Administration of Surveying, Mapping and Geoinformation announced on Monday, that the tectonic shift induced by the earthquake had moved the mountain by 1.5 inches to the southwest.
At first glance this movement doesn't seem like much at all. Look at the Napa Earthquake of August last year, where the ground shifted more than ten times as much as Mt. Everest drifted. The quake in the southern wine country caused a shift of the ground at the surface of up to 18 inches, even though it had a magnitude of "only" 6.0 and was therefore almost 900 times less energetic than the Nepal quake. During the Great San Francisco Earthquake of 1906 the ground moved a whopping 18 feet, even though its magnitude was supposedly very similar to the Nepal quake. Why is it that Mt. Everest moved so little, while the Point Reyes peninsula, where you can still see traces of the 1906 quake, moved so far?
|This curb in Hollister has shifted to the right, due to the steady creep of the Calaveras Fault (Photo by: Horst Rademacher)|
The main difference, of course, is the distance from the epicenter. The fence line at Point Reyes, which still shows the shift from the quake almost 110 years ago, lies directly on the San Andreas Fault, the "host" of the 1906 quake. Similarly, the broken ground in the wine country actually is the surface trace of the West Napa Fault, on which the Napa quake occurred. In contrast, the distance between Mount Everest and the epicenter of the Nepal quake is more than 140 miles - which puts the small movement of Everest into perspective. If an earthquake is capable of moving mountains at a distance of 140 miles by even a few inches, it must have been very powerful.
Such movement, what we seismologists call co-seismic slip, is common to almost all earthquakes. It is a clear indication at the Earth's surface that the two flanks of a fault have actually shifted deep below. For small quakes, the shift may be only a few hundredths of an inch - and hence hardly measureable. But it may also be huge. Take the Tohoku-Oki Earthquake off the coast of Japan in March 2011. With a magnitude of 9.1 it was one of the most powerful quakes so far this century. The closest land to its epicenter, the Oshika Peninsula of Eastern Honshu, was 43 miles away - but it moved by almost 5 yards, drawn towards the east by the rupture of this giant quake. (see figure above).
As the name suggests, such co-seismic slip occurs during the earthquake. It shall therefore not be confused with the "aseismic" slip which some faults exhibit. One can see this type of slip for instance along the Calaveras Fault, which runs through the small town of Hollister far south of the San Francisco Bay (see image on the left). There curbs along side walks are shifted and retaining walls have bends. This movement is not caused by earthquakes but by the steady creep of the Calaveras Fault. In its southern segment, this fault produces hardly any earthquakes at all. Instead it releases the tectonic stress it gathers from the movement of the Pacific Plate by gently slipping along with a speed of a few tenths of inches per year. (hra108)