introduction to Forensic Science: CJ_330
rebdeoilw24
In te
r n
al e
n er
g y
Plate Tectonics And Earthquakes 4
A bad earthquake at once destroys our oldest associations: the earth, the very emblem of solidity, has moved beneath our feet like a thin crust over a fluid;—one second of time has created in the mind a strange idea of insecurity, which hours of reflection would not have produced.
—Charles Darwin , 1835, notes for The Voyage of the Beagle
CHAPTER
Photo by Peter W. Weigand.
LEARNING OUTCOMES
Movements along tectonic-plate edges are responsible for many large earthquakes. After studying this chapter, you should:
• be able to describe the types of movements along tectonic-plate edges and the resultant earthquake magnitudes.
• be able to explain why subduction-zone earthquakes have the greatest magnitudes.
• understand the seismic-gap method of forecasting earthquakes.
• recognize the relationship between buildings and earthquake fatalities.
OUTLINE
• Tectonic-Plate Edges and Earthquakes
• Spreading-Center Earthquakes
• Convergent Zones and Earthquakes
• Subduction-Zone Earthquakes
• Continent-Continent Collision Earthquakes
• The Arabian Plate
• Transform-Fault Earthquakes
During the Northridge earthquake, the ground moved rapidly to the north and pulled out from under this elevated apartment building, causing it to fall back onto its parking lot in Canoga Park, California, 17 January 1994.
abb22878_ch04_079-109.indd 79abb22878_ch04_079-109.indd 79 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
80 Chapter 4 Plate Tectonics and Earthquakes
together—such as India slamming into Asia to uplift the Himalayas—involve incredible amounts of energy. This results in Earth’s greatest earthquakes.
Moving from an idealized plate, let’s examine an actual plate—the Pacific plate. Figure 4.2 shows the same type of plate-edge processes and expected earthquakes. The Pacific plate is created at the spreading centers along its eastern and southern edges. The action there produces smaller earthquakes that also happen to be located away from major human populations.
The slide-past motions of long transform faults occur: (1) in the northeastern Pacific as the Queen Charlotte fault, located near a sparsely populated region of Canada; (2) along the San Andreas fault in California with its famous earthquakes; and
The past decade brought a staggering number of mega-killer earthquakes ( table 4.1 ). The causes of these earthquakes are best understood using their plate- tectonic settings.
Tectonic-Plate Edges And Earthquakes Most earthquakes are explainable based on plate-tectonics theory. The lithosphere is broken into rigid plates that move away from, past, and into other rigid plates. These global- scale processes are seen on the ground as individual faults where Earth ruptures and the two sides move past each other in earthquake-generating events.
Figure 4.1 shows an idealized tectonic plate and assesses the varying earthquake hazards that are concentrated at plate edges:
1. The divergent or pull-apart motion at spreading centers causes rocks to fail in tension. Rocks rupture relatively easily when subjected to tension. Also, much of the rock here is at a high temperature, causing early failures. Thus, the spreading process yields mainly smaller earthquakes that do not pose an especially great threat to humans.
2. The slide-past motion occurs as the rigid plates fracture and move around the curved Earth. The plates shear and slide past each other in the dominantly horizontal movements of transform faults. This process creates large earthquakes as the irregular plate boundaries retard movement because of irregularities along the faults. It takes a lot of stored energy to overcome the rough surfaces, nonslippery rocks, and bends in faults. When these impediments are finally over- come, a large amount of seismic energy is released .
3. The convergent or push-together motions at subduction zones and in continent-continent collisions cause rocks to fail in compression. These settings store immense amounts of energy that are released in Earth’s largest tectonic earthquakes. The very processes of pulling a 70 to 100 km (45 to 60 mi) thick oceanic plate back into the mantle via a subduction zone or of pushing continents
TRANSFORM FAULT
TRANSFORM FAULT
Plate movement
Larger earthquakes
Slide-past motion
Gig ant
ic e art
hqu ake
s
Ov err
idin g p
late
Do wn
goi ng
pla te
SU BD
UC TIO
N Z ON
E
Co llid
ing mo
tion
Co nve
rge nce
zo ne
SP RE
AD ING
CE NT
ER
Pu ll-a
par t m
otio n
Div erg
enc e z
one
Sm alle
r e art
hqu ake
s
Figure 4.1 Map view of an idealized plate and the earthquake potential along its edges.
Mega-Killer Earthquakes, 2003–2011
Year Place Magnitude Deaths Tectonic Setting 2011 Japan 9.0 ̃22,000 subduction
2010 Haiti 7.0 ̃230,000 transform fault
2008 China 7.9 87,500 continent collision
2005 Pakistan 7.6 88,000 continent collision
2004 Indonesia 9.1 ̃245,000 subduction
2003 Iran 6.6 31,000 continent collision
TABLE 4.1
abb22878_ch04_079-109.indd 80abb22878_ch04_079-109.indd 80 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Spreading-Center Earthquakes 81
(3) at the southwestern edge of the Pacific Ocean where the Alpine fault cuts across the South Island of New Zealand (see figures 3.5 and 3.6).
The Pacific plate subducts along its northern and western edges and creates enormous earthquakes, such as the 2011 Japan seism, the 1964 Alaska event, and the 1931 Napier quake on the North Island of New Zealand.
Our main emphasis here is to understand plate-edge effects as a means of forecasting where earthquakes are likely to occur and what their relative sizes may be.
Spreading-Center Earthquakes A look at earthquake epicenter locations around the world (see figure 2.20) reveals that earthquakes are not as common in the vicinity of spreading centers or divergence zones as they are at transform faults and at subduction/collision zones. The expanded volumes of warm rock in the oceanic ridge systems have a higher heat content and a resultant decrease in rigidity. These heat-weakened rocks do not build up and store the huge stresses necessary to create great earthquakes.
ICELAND The style of spreading-center earthquakes can be appreciated by looking at the earthquake history of Iceland, a nation that exists solely on a hot-spot–fed volcanic island portion of the mid-Atlantic ridge spreading center ( figures 4.3 and 4.4 ). The Icelandic geologist R. Stefansson reported on catastrophic earthquakes in Iceland and stated that in the
0° 0°
Pacific plate
San Andreas fault
Queen Charlotte fault
Alpine fault
Figure 4.2 The Pacific plate is the largest in the world; it underlies part of the Pacific Ocean. Its eastern and southern edges are mostly spreading centers characterized by small- to intermediate-size earthquakes. Three long transform faults exist along its sides in Canada (Queen Charlotte), California (San Andreas), and New Zealand (Alpine); all are characterized by large earthquakes. Subduction zones (shown by black triangles) lie along the northern and western edges, from Alaska to Russia to Japan to the Philippines to Indonesia to New Zealand; all are characterized by gigantic earthquakes. From P. J. Wyllie, The Way the Earth Works. Copyright © 1976 John Wiley & Sons, Inc., New York. Reprinted with permission of John Wiley & Sons, Inc.
portions of the country underlain by north-south-oriented spreading centers, stresses build up to cause earthquakes too small to destroy buildings or kill people. These moderate- size earthquakes tend to occur in swarms, as is typical of volcanic areas where magma is on the move. Iceland does have large earthquakes, but they are associated with
Eurasian plate
North American plate
Iceland
Atlantic Ocean
Thingvellir M
id -A
tla nt
ic
rid ge
M id
-A
tla nt
ic rid
ge
Figure 4.3 Iceland sits on top of a hot spot and is being pulled apart by the spreading center in the Atlantic Ocean. Triangles mark sites of some active volcanoes.
abb22878_ch04_079-109.indd 81abb22878_ch04_079-109.indd 81 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
82 Chapter 4 Plate Tectonics and Earthquakes
east-west-oriented transform faults between the spreading- center segments.
RED SEA AND GULF OF ADEN Iceland has been built on a mature spreading center that has been opening the North Atlantic Ocean basin for about 180 million years. What would a young spreading center and new ocean basin look like? Long and narrow. In today’s world, long and narrow ocean basins exist in northeast Africa as the Red Sea and the Gulf of Aden ( figure 4.5 ). Following is a model explaining how spread- ing began: the northeastern portion of Africa sits above an extra-hot area in the upper mantle. The heat contained within this mantle hot zone is partially trapped by the blanketing effect of the overlying African plate and its embedded continent ( figure 4.6 a). The hot rock expands in volume, and some liquefies to magma. This volume expan- sion causes doming of the overlying rocks, with resultant uplift of the surface to form topography ( figure 4.6 b). The doming uplift sets the stage for gravity to pull the raised landmasses downward and apart, thus creating pull-apart faults with centrally located, down-dropped rift valleys, also described as pull-apart basins ( figure 4.6 c). As the fracturing/faulting progresses, magma rises up through the cracks to build volcanoes. As rifting and volcanism continue, seafloor spreading processes take over, the down-dropped linear rift valley becomes filled by the ocean, and a new sea is born ( figure 4.6 d).
Figure 4.5 reveals another interesting geometric feature. Three linear pull-apart basins meet at the south end of the Red Sea; this point where three plate edges touch is called a triple junction. Three rifts joining at a point may concen- trate mantle heat, or a concentration of heat in the upper mantle may begin the process of creating this triple junction. Earth’s surface may bulge upward into a dome, causing the elevated rocks to fracture into a radial pattern ( figure 4.7 ). Gravity can then pull the dome apart, allowing magma to well up and fill three major fracture zones, and the spreading process is initiated.
The triple junction in northeast Africa is geologically young, having begun about 25 million years ago. To date, spreading in the Red Sea and Gulf of Aden has been enough to split off northeast Africa and create an Arabian plate and to allow seawater to flood between them. But the East African Rift Valley has not yet been pulled far enough apart for the sea to fill it (see figure 4.5 ). The East African Rift Valley is a truly impressive physiographic feature. It is 5,600 km (3,500 mi) long and has steep escarpments and dramatic val- leys. Beginning at the Afar triangle at its northern end and moving southwest are the domed and stretched highlands of Ethiopia, beyond which the Rift Valley divides into two major branches. The western rift is markedly curved and has many deep lakes, including the world’s second deepest lake, Lake Tanganyika. The eastern rift is straighter and holds shallow, alkaline lakes and volcanic peaks, such as Mount
Baghdad
Cairo
N ile
R iv
er
Ea st
A fri
ca n
R ift
V al
le y
R ed S
ea
Black Sea
Persian Gulf
Caspian Sea
Gul f of
Ade n
Afar Triangle
Lake Albert
Lake Victoria
Lake Turkana
Nairobi
Aral Sea
Zagros Mtns
Caucasus Mtns
African plate
Arabian plate
Somali plate
Figure 4.5 Topography in northeastern Africa and Arabia. Northeastern Africa is being torn apart by three spreading centers: Red Sea, Gulf of Aden, and East African Rift Valley. The spreading centers meet at the triple junction in the Afar Triangle.
Figure 4.4 Looking south along the fissure at Thingvellir, Iceland. This is the rift valley being pulled apart in an east-west direction by the continuing spreading of the Atlantic Ocean basin. Photo by John S. Shelton.
abb22878_ch04_079-109.indd 82abb22878_ch04_079-109.indd 82 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Spreading-Center Earthquakes 83
Asthenosphere
Melting
Oceanic crust melting
Magma
Expanding
Hot rocks
Hot rocks
(a) Stage 1, Centering
(b) Stage 2, Doming
(c) Stage 3, Rifting
(d) Stage 4, Spreading
Lithosphere
Extension due to heating
Hot region in mantle
Figure 4.6 A model of the stages in the formation of an ocean basin. (a) Stage 1, Centering: Moving lithosphere centers over an especially hot region of the mantle. (b) Stage 2, Doming: Mantle heat causes melting, and the overlying lithosphere/continent extends. The increase in heat causes surface doming through uplifting, stretching, and fracturing. (c) Stage 3, Rifting: Volume expansion causes gravity to pull the uplifted area apart; fractures fail and form faults. Fractures/ faults provide escape for magma; volcanism is common. Then, the dome’s central area sags downward, forming a valley such as the present East African Rift Valley. (d) Stage 4, Spreading: Pulling apart has advanced, forming a new seafloor. Most magmatic activity is seafloor spreading, as in the Red Sea and the Gulf of Aden.
abb22878_ch04_079-109.indd 83abb22878_ch04_079-109.indd 83 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
84 Chapter 4 Plate Tectonics and Earthquakes
Kilimanjaro, Africa’s highest mountain. The Rift Valley holds the oldest humanoid fossils found to date and is the probable homeland of the first human beings. Will the spreading continue far enough to split a Somali plate from Africa? It is simply too early to tell.
How severe are the earthquakes in the geologically youthful Red Sea and Gulf of Aden? Moderately—but these spreading-center earthquakes are not as large as the earth- quakes on the other types of plate edges.
GULF OF CALIFORNIA The Red Sea and Gulf of Aden spreading centers of the Old World have analogues in the New World with the spreading centers that are opening the Gulf of California and moving the San Andreas fault ( figure 4.8 ). Geologically, the Gulf of California basin does not stop where the sea does at the northern shoreline within Mexico. The opening ocean basin continues northward into the United States and includes the Salton Sea and the Imperial and Coachella valleys at the ends of the Salton Sea. The Imperial Valley region is the only part of the United States that sits on opening ocean floor. In the geologic past, this region was flooded by the sea. However, at present, fault movements plus the huge volume of sedi- ment deposited by the Colorado River hold back the waters of the Gulf of California. If the natural dam is breached, the United States will trade one of its most productive agricul- tural areas for a new inland sea.
The spreading-center segment at the southern end of the Salton Sea is marked by high heat flow, glassy volcanic domes, boiling mud pots, major geothermal energy reser- voirs (subsurface water heated to nearly 400°C (750°F) by the magma below the surface), and swarms of earthquakes associated with moving magma ( figure 4.8 ).
Figure 4.7 Schematic map of a triple junction formed by three young spreading centers. Heat may concentrate in the mantle and rise in a magma plume, doming the overlying lithosphere and causing fracturing into a radial set with three rifts. Gravity may then pull the dome apart, initiating spreading in each rift.
Gravitational Spreading
Sp re
ad in
g
Spreading
pull
Basin and
Range
C ol
or ad
o R
iv er
fault
fault
Prieto
fault
Elsinore
San Andreas fault
San Jacinto fault
Salton Sea Laguna Salada fault
Laguna
Salada
Peninsular Ranges
Gulf of California
Geothermal area Volcanic dome
California
Baja California Mexicali
Valley
Imperial Valley
Coachella Valley
Im perial
Cerro
ArizonaSonora
N
Figure 4.8 Map of northernmost Gulf of California. Note the two spreading centers (shown in red and by large, diverging arrows) and the right-lateral (transform) faults associated with them.
The Salton Trough is one of the most earthquake-active areas in the United States. There are seisms caused by the splitting and rifting of continental rock and swarms of earth- quakes caused by forcefully moving magma. The Brawley seismic zone at the southern end of the Salton Sea commonly experiences hundreds of earthquakes in a several-day period. For example, in four days in January 1975, there occurred 339 seisms with magnitudes (M L ) greater than 1.5; of these, 75 were greater than 3M L , with the largest tremor at 4.7M L . Because hot rock does not store stress effectively, energy release takes place via many smaller quakes. The larger earthquakes in the valley are generated by ruptures in brittle continental rocks.
Notice in figure 4.8 that the San Andreas fault ends at the southeastern end of the Salton Sea at the northern limit of the spreading center. Notice also that other major faults, such as the Imperial, San Jacinto system, Cerro Prieto, Elsinore, and Laguna Salada, also appear to be transform faults that line up with spreading-center segments. From a broad perspective, all these subparallel, right-lateral, transform faults are part of
abb22878_ch04_079-109.indd 84abb22878_ch04_079-109.indd 84 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Subduction-Zone Earthquakes 85
mantle, whereas continents float about on the asthenosphere in perpetuity. Continents are ripped asunder and then reassembled into new configurations via collisions, but they are not destroyed by subduction.
Subduction-Zone Earthquakes Subduction zones are the sites of great earthquakes. Imagine pulling a 100 km (62 mi) thick rigid plate into the weaker, deformable rocks of the mantle that resist the plate’s intru- sion. This process creates tremendous stores of energy, which are released periodically as great earthquakes. Most of the really large earthquakes in the world are due to subduction ( table 4.2 ). Subduction occurs on a massive scale. At the present rates of subduction, oceanic plates with an area equivalent to the entire surface area of Earth will be pulled into the mantle in only 180 million years.
A descending slab of oceanic lithosphere is defined by an inclined plane of deep earthquakes or fault-rupture loca- tions (see figure 2.21). Earthquakes at subduction zones result from different types of fault movements in shallow versus deeper realms. At shallow depths (less than 100 km, or 62 mi), the two rigid lithospheric plates are pushing against each other. Earthquakes result from compressive movements where the overriding plate moves upward and the subducting plate moves downward. Pull-apart fault movements also occur near the surface within the subducting
the San Andreas plate boundary fault system carrying peninsular California to the northwest. Large earthquakes on these faults in recent years in the area covered by figure 4.8 include a 6.9M w quake on the Imperial fault in 1940, a 6.6M w event in the San Jacinto system in 1942, a 6.4M w quake on the Imperial fault in 1979, a 6.6M w quake in the San Jacinto system in 1987, and a 7.25M w seism in the Laguna Salada fault system in 2010. These are large earthquakes, but deaths and damages for each event typically were not high because the region is sparsely inhabited, most buildings are low, and the frequent shakes weed out inferior buildings.
Convergent Zones and Earthquakes The greatest earthquakes in the world occur where plates collide. Three basic classes of collisions are (1) oceanic plate versus oceanic plate, (2) oceanic plate versus continent, and (3) continent versus continent. These collisions result in either subduction or continental upheaval. If oceanic plates are involved, subduction occurs. The younger, warmer, less- dense plate edge overrides the older, colder, denser plate, which then bends downward and is pulled back into the mantle. If a continent is involved, it cannot subduct because its huge volume of low-density, high-buoyancy rocks simply cannot sink to great depth and cannot be pulled into the denser mantle rocks below. The fate of oceanic plates is destruction via subduction and reassimilation within the
Earth's Largest Earthquakes, 1904–2011
Rank Location Year M w Cause 1. Chile 1960 9.5 Subduction—Nazca plate
2. Alaska 1964 9.2 Subduction—Pacific plate
3. Indonesia 2004 9.1 Subduction—Indian plate
4. Japan 2011 9.0 Subduction—Pacific plate
5. Kamchatka 1952 9.0 Subduction—Pacific plate
6. Chile 2010 8.8 Subduction—Nazca plate
7. Ecuador 1906 8.8 Subduction—Nazca plate
8. Alaska 1965 8.7 Subduction—Pacific plate
9. Indonesia 2005 8.6 Subduction—Indian plate
10. Assam 1950 8.6 Collision—India into Asia
11. Alaska 1957 8.6 Subduction—Pacific plate
12. Indonesia 2007 8.5 Subduction—Indian plate
13. Kuril Islands 1963 8.5 Subduction—Pacific plate
14. Banda Sea 1938 8.5 Subduction—Pacific/Indian plate
15. Russia 1923 8.5 Subduction—Pacific plate
16. Chile 1922 8.5 Subduction—Nazca plate
TABLE 4.2
abb22878_ch04_079-109.indd 85abb22878_ch04_079-109.indd 85 13/12/12 8:22 PM13/12/12 8:22 PM
G A R R E T T , M E G A N 1 3 2 4 T S
86 Chapter 4 Plate Tectonics and Earthquakes
JAPAN, 2011: STUCK SEGMENTS OF SUBDUCTING PLATE At first glance, the plate tectonics of eastern Japan seem fairly simple. The Pacific plate moves northwest 8.3 cm (3.3 in) per year and dives under Japan ( figure 4.9 ). The subducting plate is slowed by friction and stress builds until there is enough to rupture individual segments of the fault, resulting in earthquakes of 7 to 8 M. This has been the pat- tern of the past several centuries. But on 11 March 2011, five adjacent segments of the fault ruptured together over an area with length more than 600 km (375 mi). Two central seg- ments had enormous movements or slips up to 50 m (165 ft), which helped produce the 9.0M w seism.
The 9.0M w mainshock was preceded two days earlier by a 7.2M w event 40 km (25 mi) away and three other seisms greater than 6M. But we don’t know how to recognize that an earthquake is a foreshock until after the mainshock occurs; thus, no warnings were given. The maximum acceleration offshore is calculated as 3g, that is, three times stronger than the pull of gravity. Acceleration of 2.7g was measured onshore in Miyagi Prefecture.
Before the earthquake, 15 years of global positioning system (GPS) data showed that the eastern edge of Japan was being dragged downward. Apparently the down-going Pacific plate was stuck to the overlying continental plate and was warping it downward. The strain of this movement accumu- lated during many centuries. When the stuck zone ruptured as the Pacific plate moved downward, the overlying plate carry- ing Japan sprang upward and released elastic strain.
The Japanese have kept the best historic records of earthquakes and tsunami in the world. A look farther back in their records shows that a similar-size earthquake and tsu- nami hit the same area of northeast Japan on 13 July 869 ce. The 2011 event surprised many experts who had disregarded the older historic record. But 1,142 years is a short time in geologic history.
A concern now is that the March 2011 earthquake trans- ferred stress southward, closer to Tokyo, which was hit hard in 1923. In November 2011, an official Japan earthquake assessment committee forecast a 30% chance of a 9M event
plate as it is bent downward and snaps in tensional failure and as the overriding plate is lifted up from below. Notice in figure 2.21 that the shallow earthquakes occur (1) in the upper portion of the down-going plate, (2) at the bend in the subducting plate, and (3) in the overriding plate.
Compare the locations of the shallow earthquake sites to those of intermediate and deep earthquakes (see figure 2.20). At depths below 100 km, earthquakes occur almost exclu- sively in the interior of the colder oceanic lithosphere, the heart of the subducting slab. The high temperatures of rock in the upper mantle cause it to yield more readily to stresses and thus not build up the stored energy necessary for gigantic earthquakes. At depth, the upper and lower surfaces of the subducting slabs are too warm to generate large earthquakes. Thus, the earthquakes occur in the cooler interior area of rigid rock, where stress is stored as gravity pulls against the mantle resistance to slab penetration. In the areas of most rapid subduction, the down-going slab may remain rigid enough to spawn large earthquakes to depths in excess of 700 km (435 mi). A great earthquake that occurs deep below the surface has much of its seismic energy dissipated while traveling to the surface. Thus, the biggest disasters are from the great earthquakes that occur at shallow depths and con- centrate their energy on the surface.
Most of the subduction-zone earthquakes of today occur around the rim of the Pacific Ocean or the northeastern Indian Ocean. This is shown by the presence of most of the deep-ocean trenches (see figure 2.14) and by the dense con- centrations of earthquake epicenters (see figure 2.20).
When will the next large earthquake occur in the north- western Pacific Ocean region? A popular way of forecasting the locations of future earthquakes is the seismic-gap method. If segments of one fault have moved recently, then it seems reasonable to expect that the unmoved portions will move next and thus fill the gaps. Looking at figure 4.9 , where would you forecast large earthquakes to occur? It is easy to see the gaps in earthquake locations, and although seismic- gap analysis is logical, it yields only expectations, not guar- antees. One segment of a fault can move two or more times before an adjoining segment moves once. In 2011, a seismic gap was filled in Japan by a 9.0M W earthquake ( figure 4.9 ).
Eurasian plate
Alaska
Tokyo 1952
1933 2011
1952 1958
1963
1945
1952 1965
1957 1938
1964
1958
1946
1944 1923
1968 1969
Ja pa
n
Pacific plate
Figure 4.9 Brown patterns show severely shaken areas, with dates, from recent earthquakes caused by Pacific plate subduction. The 1957, 1964, and 1965 Alaska earthquakes are three of the largest in the 20th century. They were part of an earthquake cluster. Using the seismic-gap method, where are the next earthquakes most likely to occur? See the seismic gap filled in 2011.
abb22878_ch04_079-109.indd 86abb22878_ch04_079-109.indd 86 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Subduction-Zone Earthquakes 87
Was this second rupture event, just 92 days later, a continuation of the earlier earthquake? It appears that a bend or scissors- like tear in the subducting plate may have delayed the full rupture in December 2004. The history of the region sug- gested there were more big earthquakes to come.
And come they did. An earthquake cluster is under way. On 12 September 2007, there were two earthquakes, an 8.4M w followed 12 hours later by a 7.9M w ; on 20 February 2008 there was a 7.4M w ; on 30 September 2009 there was a 7.6M w ; and in 2010 there was a 7.8 M W on 6 April and a 7.7 M W on 25 October ( figure 4.10 ). And there are more seismic gaps to fill.
Another earthquake cluster consisting of a 9.2M w and several magnitude 8s occurred in the mid-20th century at the Pacific plate subduction zone along Alaska, Russia, and northern Japan (see figure 4.9 and table 4.2 ).
MEXICO CITY, 1985: LONG- DISTANCE DESTRUCTION On Thursday morning, 19 September 1985, most of the 18 million residents of Mexico City were at home, having their morning meals. At 7:17 a.m., a monstrous earthquake broke loose some 350 km (220 mi) away. Seismic waves trav- eled far to deal destructive blows to many of the 6- to 16-story buildings that are heavily occupied during the working day ( figure 4.11 ). Building collapses killed more than 9,000 people.
closer to Tokyo in the next 30 years. The fault there is closer to shore, meaning more intense shaking for buildings and less time before tsunami come ashore.
The death and destruction caused by the 2011 earth- quake are not known because many of its effects were erased by the tsunami that followed. We will pick up this story again in Chapter 8 on tsunami.
INDONESIA, 2004: ONE EARTHQUAKE TRIGGERS OTHERS The Indian-Australian plate moves obliquely toward western Indonesia at 5.3 to 5.9 cm/yr (2 to 2.3 in/yr). The enormous, ongoing collision results in subduction-caused earthquakes that are frequent and huge ( figure 4.10 ). Many of these earth- quakes send off tsunami. On 26 December 2004, a 1,500 km (930 mi) long fault rupture began as a 100 km (62 mi) long portion of the plate-tectonic boundary ruptured and slipped during 1 minute. The rupture then moved northward at 3 km/ sec (6,700 mph) for 4 minutes, then slowed to 2.5 km/sec (5,600 mph) during the next 6 minutes in a 9.1M w event. At the northern end of the rupture, the fault movement slowed drasti- cally and only traveled tens of meters during the next half hour.
On 28 March 2005, the subduction zone broke again and this time ruptured 400 km (250 mi) southward from the south- ern end of the 2004 rupture in an 8.6M w event ( figure 4.10 ).
The Tokyo Earthquake of 1923 Early on Saturday morning, 1 September 1923, the cities of Tokyo and Yokohama were shattered by a deadly series of earthquakes. The principal shock occurred as the floor of Sagami Bay dropped markedly and sent 11 m (36 ft) high seismic sea waves (tsunami) crashing against the shore. The waves washed away hundreds of homes. Yet fishermen spending their day at sea were unaware of the monster waves. At day's end, as they sailed toward home through Sagami Bay, they were sickened to find the floating wreckage of their houses and the bodies of their families. Devastation on land was great; houses were destroyed, bridges fell, tunnels col- lapsed, and landslides destroyed hills. The shaking caused the collapse of flammable house materials onto cooking fires, and the flames, once liberated, quickly raced out of control. Little could be done to stop their spread because the earthquake had broken the water mains. Shifting winds pushed the fires for two and a half days, destroying 71% of Tokyo and 100% of Yokohama.
Possibly the most tragic event in this disaster occurred when 40,000 people, clutching their personal belongings,
A Classic Disaster
attempted to escape the flames by crowding into a 250-acre garden on the edge of the Sumida River. People packed themselves into this open space so densely that they were barely able to move. At about 4 p.m., several hours after the earthquake, the roaring fires approached on all three land- ward sides of the crowd. Suddenly the fire-heated winds spawned a tornado that carried flames onto the huddled masses and their combustible belongings. After the flames had died, 38,000 people lay dead, either burned or asphyxi- ated. The usual instinct to seek open ground during a disaster was shockingly wrong this time.
The combined forces of earthquakes, tsunami, and fires killed 99,331 people and left another 43,476 missing and presumed dead. Yet, despite this immense catastrophe, the morale of the Japanese people remained high. They learned from the disaster. They have rebuilt their cities with wider streets, more open space, and less use of combustible con- struction materials.
The historic record of earthquakes in the region is thought provoking. The region 80 km (50 mi) southwest of Tokyo has been rocked by five very strong earthquakes in the last 400 years. The seisms have occurred roughly every 73 years, the most recent in 1923.
abb22878_ch04_079-109.indd 87abb22878_ch04_079-109.indd 87 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
88 Chapter 4 Plate Tectonics and Earthquakes
Guerrero seismic gap lies near Acapulco and is closer to Mexico City than the Michoacan epicenter.
Resonance Matters Many of the coastal towns near the epicenter received relatively small amounts of damage. Yet in Mexico City, more than 5,700 buildings were severely damaged, with 15% of them collapsing catastrophically. Why did so many buildings collapse and kill so many people when Mexico City lies 350 km (220 mi) from the epicenter? It was largely due to resonance between seismic waves, soft lake-sediment foundations, and improperly designed buildings. The duration of shaking was increased due to seismic energy trapped within the soft sediments.
What caused this earthquake? The Cocos plate made one of its all-too-frequent movements. This time, a 200 km (125 mi) long front, inclined 18° east, thrust downward and eastward about 2.3 m (7.5 ft) in two distinct jerks about 26 seconds apart ( figure 4.12 ). The mainshock had a surface wave magnitude (M s ) of 8.1. It was followed on 21 September by a 7.5M s aftershock and by another on 25 October of 7.3M s . The earthquakes were not a surprise to seismologists. Before these seisms occurred, the area was called the Michoacan seismic gap, and many instruments had been deployed in the region to measure the expected big event.
As figure 4.12 shows, another large seismic gap waits to be filled by a major movement of the Cocos plate. The
S u b d u c
t i o
n z o n e
S u
b d
u c
t i o
n z
o n
e
Burma plate
Indian- Australian plate
Sunda plate
Thailand
90°E 95°E 100°E 105°E
15°N
10°N
5°N
0°
5°S
Cambodia
Malaysia
5.7 cm/yr
1833 M9.0
1797 M8.8
1861 M8.5 1907 M7.8
2005 M8.6
2004 M9.2 2002 M7.3
2010 M7.8
2010 M7.7
2008 M7.4
1935 M7.7
2009 M7.6
2007 M7.9
2000 M7.9
2007 M8.4
1881 M7.9 1941 M7.7
5.3 cm/yr
Sumatra
Java
Figure 4.10 Subduction of the Indian-Australian plate beneath Indonesia was the cause of the huge earthquakes in 2004, 2005, 2007, 2008, 2009 and 2010. The region has a long history of large earthquakes, and more will occur.
abb22878_ch04_079-109.indd 88abb22878_ch04_079-109.indd 88 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Subduction-Zone Earthquakes 89
vibrated in the 1- to 2-second frequency band. Where all three factors were in phase, disaster struck.
There were design flaws in the failed buildings ( figure 4.14 ), including soft first stories, poorly joined building wings, odd-shaped buildings prone to twist on their foundations, and buildings of different heights and vibration frequencies that sat close together and bumped into each other during the earthquake ( figures 4.14 c and 4.15 ).
CHILE, 1960: THE BIGGEST ONE Earth’s biggest measured earthquake is a 9.5M w event that sprang forth on Sunday afternoon, 22 May 1960, in southern Chile. Here the Nazca plate converges with the South American plate at 8 m/century (see figure 2.14). In 1960, the
Mexico City is built atop the former Aztec capital of Tenochtitlan. The Aztecs built where they saw the favorable omen—an eagle sitting on a cactus and holding a writhing snake in its mouth. The site was Lake Texcoco, a broad lake surrounded by hard volcanic rock. Over time, the lake basin was partially filled with soft, water-saturated clays. Portions of Lake Texcoco have been drained, and large buildings have been constructed on the weak lake-floor sediments.
Building damages were the greatest and the number of deaths the highest where three factors combined and created resonance: (1) the earthquakes sent a tremendous amount of energy in seismic waves in the 1- to 2-second frequency band; (2) the areas underlain by thick, soft muds (clays) vibrating at 1- to 2-second frequencies amplified the seismic waves ( figure 4.13 ); and (3) buildings of 6 to 16 stories
*
*
28 Jul y 57
7 Ju
ne 62
21 Sept 85
19 Sept 85
30 Jan 73
Oaxac
Guerrero state
Mexico CityMichoacan stateNorth
American plate (overriding)
400 km
Shoreline
Guerreroseismic gap
Cocos plate (subducting)
Pacific Ocean
16°
17°
18°
19°N
104°W 103° 102° 101° 100° 99° 98°
Acapulco Filled
Michoacan gap
state
14 Mar
79
East–west acceleration
UNAM on hard rocks
SCT on soft muds
20 0
cm /s
ec 2
Time (seconds)
0
0 10 20 30 40 50 60
0
Figure 4.11 This 15-story building collapsed during the 1985 Mexico City earthquake, crushing all its occupants as its concrete floors pancaked. Photo by M. Celebi, US Geological Survey .
Figure 4.12 Map of coastal Mexico showing dates of earthquakes and fault areas moved (dashed lines) during Cocos plate subduction events. The Michoacan seismic gap was filled by the 1985 seisms. The Guerrero seismic gap is overdue for a major movement.
Figure 4.13 Some east–west accelerations recorded in Mexico City in 1985. The Universidad Nacional Autonoma de Mexico (UNAM) sits on a hard-rock hill and received small accelerations. The Secretaria de Comunicaciones y Transportes site (SCT) sits on soft lake sediments that amplified the seismic waves.
abb22878_ch04_079-109.indd 89abb22878_ch04_079-109.indd 89 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
90 Chapter 4 Plate Tectonics and Earthquakes
Nazca plate moved eastward and downward while the South American plate moved westward and above with slips of 20 to 30 m (65 to 100 ft). In the 33 hours before the big one, there were foreshocks of 8.1M w and six others of greater than 6M. Over a period of days, the subduction-zone ruptures
involved a 1,000 km (620 mi) length and a 300 km (185 mi) width.
Chile, 1835 and 2010: Emptying, Filling, and Emptying Elastic Strain in a Seismic Gap Giant earthquakes are common phenomena in Chile. During his epic voyage on the HMS Beagle , Charles Darwin was resting on his back in the woods near Valdivia, Chile, on 20 February 1835 when a huge earthquake struck. His well- written description of large areas of land being uplifted, giant sea waves crossing the shoreline, and two volcanoes being shaken into eruptions are instructive even today. The modern estimate is that the elastic strain released in Darwin’s earth- quake yielded an 8.5M w event.
During the 175 years following the 1835 seism, tectonic- plate convergence in the region was about 14 m (46 ft), but few earthquakes were being recorded. A seismic gap was recog- nized and the region was heavily instrumented. GPS measure- ments showed that the area was not moving; it was locked. Frictional resistance was causing elastic strain to accumulate in the rocks; the seismic gap was filling with strain. A big earthquake was expected, and on 27 February 2010, it hap- pened: the 8.8M w event is the 6
th biggest earthquake ever measured. A 500 km (310 mi) long subduction-zone interface ruptured with bilateral movement at 3.1 km/sec (6,930 mph). Some areas of fault surface moved up to 15 m (50 ft), whereas other areas had low to no slip. The elastic-strain buildup in the seismic gap was largely emptied. South America, from Chile to the Argentina coast, had all moved westward.
ALASKA, 1964: SECOND BIGGEST ONE Saint Matthew’s account of the first Good Friday included: “And, behold . . . the earth did quake, and the rocks rent.” His words applied again, more than 1,900 years later, on Good Friday, 27 March 1964. At 5:36 p.m., in the wilderness at the head of Prince William Sound, a major subduction move- ment created a gigantic earthquake. This was followed in sequence by other downward thrusts at 9, 19, 28, 29, 44, and 72 seconds later as a nearly 1,000 km (more than 600 mi) long slab of 400 km (250 mi) width lurched its way deeper into the mantle. Hypocenter depths were from 20 to 50 km (12 to 30 mi). The earthquake magnitude was 9.2M w .
The duration of strong ground shaking was lengthy— 7 minutes; it induced many avalanches, landslides, ground settlements, and tsunami. Of the 131 lives lost, 122 were due to tsunami. The town of Valdez was severely damaged by both ground deformation and a submarine landslide that caused tsunami, which destroyed the waterfront facilities. Damage was so great that the town was rebuilt at a new site. Anchorage, the largest city in Alaska, was heavily damaged by landslides. And yet, there were some elements of luck in the timing of this earthquake. It occurred late on a Friday, when few people were in heavily damaged down- town Anchorage; tides were low; it was the off-season for
(c) (d)
(b) Ground movement(a)
Figure 4.14 Some building-response problems during the Mexico City earthquake. (a) The amplitude of shaking increases up the building. (b) Buildings with long axes perpendicular to ground motion suffer more shaking. (c) Buildings with different heights sway at different frequencies and bang into each other. (d) A building with different heights tends to break apart.
Figure 4.15 Mexico City earthquake damage caused by constructing buildings with different periods of vibration next to each other. The four-story building on the left repeatedly struck the taller Hotel de Carlo (middle building), causing collapse of its middle floors (see figure 4.14 c). The taller building on the right was damaged by hammering from the Hotel de Carlo. Photo from NOAA.
abb22878_ch04_079-109.indd 90abb22878_ch04_079-109.indd 90 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Subduction-Zone Earthquakes 91
giant earthquakes? Could the Cascadia subduction zone be plugged up like a clogged drain, meaning that subduc- tion has stopped? No. The active volca- noes above the subducting plates testify that subduction is still occurring. Could the subduction be taking place smoothly and thus eliminating the need for giant earthquakes? Probably not. The oceanic lithosphere being sub- ducted is young, only about 10 million years old. Young lithosphere is more buoyant and is best subducted when overridden by continental lithosphere. The North American continent is mov- ing southwest at 2.5 cm/yr (1 in/yr) and colliding with the oceanic plate, which is subducting along a N 68° E path at 3.5 cm/yr (1.4 in/yr). Thus, it seems certain that the subduction zone is stor- ing energy in elastic strain.
The Cascadia subduction zone is 1,100 km (680 mi) long. Its character- istics of youthful oceanic plate and strong coupling with the overriding plate are similar to situations in south-
western Japan and southern Chile. Events of Chilean magnitude could unlock the entire Cascadia subduction
zone. Figure 4.16 is a plot of the epicenters of the 1960 Chile mainshock, foreshocks, and aftershocks over a map of the Pacific Northwest to give an idea of what could happen in British Columbia, Washington, and Oregon. Could the Pacific Northwest experience a magnitude 9 earthquake? Yes—in fact, it already has.
Earthquake in 1700: The Trees Tell the Story Recent work by Brian Atwater has shown that the last major earthquake in the Pacific Northwest occurred about 9 p.m. on 26 January 1700 and was about magnitude 9. This is indicated by two converging lines of evidence: (1) Counting the annual growth rings in trees of drowned forests along the Oregon–Washington–British Columbia coast shows that the dead trees have no rings after 1699. Apparently the ground dropped during an earthquake, and seawater got to the tree roots, killing them between August 1699 and May 1700, the end of one growing season and the beginning of the next ( figure 4.17 ). (2) The Japanese maintain detailed records of tsunami occurrences and sizes that they correlate to earthquake magnitudes and locations around the Pacific Ocean. Tsunami of 2 m (7 ft) height that hit Japan from midnight to dawn point to a 9 p.m. earthquake along the Washington–Oregon coast on 26 January 1700.
What will the British Columbia–Washington–Oregon region experience during a magnitude 9 earthquake? Three to five minutes of violent ground shaking will be followed by
Figure 4.16 Epicenters for the 1960 Chile earthquake sequence are plotted over the Cascadia subduction zone. One earthquake had a magnitude of 9.5; nine had magnitudes of 7 to 8; and 28 had magnitudes of 6 to 7.
Pacific plate
Juan de Fuca plate
Gorda plate
Explorer plate Cascadia subduction zone
Epicenters 9.5 magnitude 7 to 8 magnitude 6 to 7 magnitude
132° W 116° W 51° N
British Columbia
California Nevada
200 km
120 mi
Oregon
Washington
Vancouver
Longview
Portland Salem
Corvallis
Roseburg
Medford
Canada USA
Seattle Tacoma Olympia
200 km
120 mi 40° N
0
0
Eugene
0
0
fishing, so few people were on the docks or in the canner- ies; and the weather in the ensuing days was seasonally warm, thus sparing people from death-dealing cold while their homes and heating systems were out of order.
If Alaska had been a densely inhabited area, the dimen- sions of the human catastrophe would have been mind-boggling. In 2004, essentially the same-size earthquake and tsunami occurred in Indonesia, killing at least 245,000 people in the region, compared to 131 in the Alaska event.
In the United States, California is commonly called “earthquake country,” but it is clear from table 4.2 that Alaska is more deserving of this title. Over the past 5 million years, about 290 km (180 mi) of Pacific plate have been pulled under southern Alaska in the vicinity of Anchorage.
PACIFIC NORTHWEST: THE UPCOMING EARTHQUAKE The 1985 Mexico City earthquake was caused by eastward subduction of a small plate beneath the North American plate. Other small plates are subducting beneath North America at the Cascadia subduction zone ( figure 4.16 ). No gigantic seisms have occurred in the Pacific Northwest in the 200 or so years since Europeans settled there. Will this area remain free of
abb22878_ch04_079-109.indd 91abb22878_ch04_079-109.indd 91 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
92 Chapter 4 Plate Tectonics and Earthquakes
tsunami 10 m (33 ft) high surging onshore 15 to 40 minutes after the earthquake. Energy will be concentrated in long- period seismic waves, presenting challenges for tall build- ings and long bridges.
What will the next magnitude 9 earthquake, along with its major aftershocks, do to cities such as Portland, Tacoma, Seattle, Vancouver, and Victoria? When will the next magni- tude 9 earthquake occur in the Pacific Northwest?
Continent-Continent Collision Earthquakes The grandest continental pushing match in the modern world is the ongoing ramming of Asia by India. When Gondwana- land began its breakup, India moved northward toward Asia. The 5,000 km (3,000 mi) of seafloor (oceanic plate) that lay in front of India’s northward path had all subducted beneath Asia by about 40 million years ago. Then, with no seafloor left to separate them, India punched into the exposed under- belly of Asia ( figure 4.18 ). Since the initial contact, the assault has remained continuous. India has moved another 2,000 km (1,250 mi) farther north, causing complex accom- modations within the two plates as they shove into, under, and through each other accompanied by folding, overriding, and stacking of the two continents into the huge mass of the Himalayas and the Tibetan Plateau. The precollision crusts of India and Asia were each about 35 km (22 mi) thick. Now, after the collision, the combined crust has been thickened to 70 km (44 mi) to create the highest-standing continental area on Earth. The Tibetan Plateau dwarfs all other high land- masses. In an area the size of France, the average elevation exceeds 5,000 m (16,400 ft). But what does all of this have
Living tree Pattern of annual growth rings
Year 1075
Year 1320
Dead tree High tide levels
Before EQ
After EQ
1699
2000
Figure 4.17 Annual growth rings in drowned trees along the Oregon–Washington–British Columbia coast tell of their deaths after the 1699 growing season. Seawater flooding occurred as land dropped during a magnitude 9 earthquake (EQ).
Sri Lanka
Sri Lanka
71 million years ago
55 million years ago
38 million years ago
10 million years ago
"INDIA" Landmass
Eurasian plate
Indian Ocean
INDIA Today
Equator
Figure 4.18 Map showing the movement of India during the past 71 million years. India continues to shove into Eurasia, creating great earthquakes all the way through China.
to do with earthquakes? Each year, India continues to move about 5 cm (2 in) into Asia along a 2,000 km (1,250 mi) front. This ongoing collision jars a gigantic area with great earthquakes. The affected area includes India, Pakistan, Afghanistan, the Tibetan Plateau, much of eastern Russia, Mongolia, and most of China.
A relatively simple experiment shows how earthquake-generating faults may be caused by continental collision ( figure 4.19 ). The experiment uses a horizontal jack to push into a pile of plasticine, deforming it under the force. The experimental defor- mation is similar to the tectonic map of the India-Asia region ( figure 4.20 ). The northward wedging of India seems to be forcing Indochina to escape to the southeast and is driving a large block of China to the east.
abb22878_ch04_079-109.indd 92abb22878_ch04_079-109.indd 92 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Continent-Continent Collision Earthquakes 93
the central Himalaya front that has not moved since 1505. The rapid population growth in these countries has resulted in the construction of millions of new buildings. Many build- ings were and are being built without seismic-safety inspec- tions to guide their construction. Even where codes exist, poor construction practices have led to catastrophic failures of many buildings during shaking.
The earthquakes of recent years have been deadly, but none of them have been a direct hit on the mega-cities of the region. Some Indian plate–caused earthquakes in China show how disheartening the death totals can be.
CHINA, 1556: THE DEADLIEST EARTHQUAKE In Shaanxi Province, the deadliest earthquake in history occurred in 1556 when about 830,000 Chinese were killed in and near Xi’ an on the banks of the mighty Huang River (once known as the Yellow River). The region has numerous hills composed of deposits of windblown silt and fine sand that have very little cohesion (ability to stick together). Because of the ease of digging in these loose sediments, a tremendous number of the homes in the region were caves dug by the
Figure 4.19 Simulated collision of India into Asia. A wedge is slowly jacked into layered plasticine confined on its left side but free to move to the right. From top to bottom of figure, notice the major faults that form and the masses that are compelled to move to the right. Compare this pattern to the tectonic map of India and Asia in figure 4.20 . After P. Tapponier, et al. (1982). Geology, 10, 611–16.
Tangshan
ShaanxiSichuan 2008Tibetan
Plateau Quetta 1935
Kashmir 2005
Rann of Kutch 1819
Bhuj 2001
Calcutta
Assam
km 500
Spreading center
Horizontal movement on fault
Subduction zone upper plate
Compressive fault overriding side
0 0
mi 310
1556
1976
H i m a l a y a
CHINA, PAKISTAN, AND INDIA, 2008, 2005, AND 2001: CONTINENT COLLISION KILLS India’s continuing push into Asia has caused three deadly earthquakes in the 21st century ( figure 4.20 ). The May 2008 Sichuan, China, event killed about 87,500 people; the Octo- ber 2005 Kashmir, Pakistan, earthquake killed about 88,000 people; and the January 2001 Gujarat, India, event killed more than 20,000. These earthquakes were close together in space and time. Is this just a coincidence, or will they be part of a cluster of ongoing killer events? We don’t know. But there are seismic gaps waiting to be filled in this region, and some are large, including a 600 km (375 mi) long section in
Figure 4.20 Tectonic map showing India pushing into Asia. The ongoing collision causes devastating earthquakes, each killing tens or hundreds of thousands of people. The list includes two in the Indian state of Gujarat (in 1819 at Rann of Kutch and in 2001 near Bhuj), three in China (in 1556 at Shaanxi, in 1976 at Tangshan, and in 2008 at Sichuan), and two in Pakistan (in Quetta in 1935 and in Kashmir in 2005).
abb22878_ch04_079-109.indd 93abb22878_ch04_079-109.indd 93 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
94 Chapter 4 Plate Tectonics and Earthquakes
Iran, 1962–2011: Mud-Block Buildings Kill Catastrophic earthquakes occur along the entire length of Iran. In the past 50 years, 10 earthquakes have killed more than 150,000 people (table 4.3). For example, the deadliest natural disaster in 2003 was the earthquake that shook loose 8 km (5 mi) below the city of Bam at 5:27 a.m. on Friday, 26 December. Traditional construction methods there con- sist of sun-dried, mud-block walls topped by heavy roofs. When the earthquake shaking began, the walls crumbled and the heavy roofs crashed down. Collapsing homes killed 41,000 people.
TRANSFORM-FAULT EARTHQUAKES The Arabian plate is moving away from Africa and is push- ing into Eurasia, but what is happening along the sides of the Arabian plate? The slide-past movements of transform faults. On the eastern side, the plate-boundary fault occurs beneath the Indian Ocean and has scant effect on humans. But look where the slide-past fault movements occur along the western side of the Arabian plate (figures 4.21, 4.22, and 4.23).
Dead Sea Fault Zone: Holy Land Disasters The Dead Sea fault zone is an Eastern Hemisphere analogue of the San Andreas fault in California. It not only runs right through the Holy Land but has also created much of the area’s well-known topography. Notice in figure 4.22 that there are four prominent overlaps or steps in the Dead Sea fault zone. Fault movements on both sides of these steps have created pull-apart basins that are filled by historically famous water bodies, such as the Dead Sea and the Sea of Galilee. The Dead Sea fault zone has been operating for as
inhabitants. Most of the residents were in their cave homes at 5 a.m. on the wintry morning of 23 January, when the seismic waves rolled in from the great earthquake. The severe shaking caused many of the soft silt and sand sediments of the region to vibrate apart and literally behave like fluids. Most of the cave-home dwellers were entombed when the once-solid walls of their homes liquefied and collapsed.
The Arabian Plate The emergence of the geologically young spreading centers in the Red Sea and Gulf of Aden has cut off the northeast tip of the African continent ( figure 4.21 ; see figure 4.5 ) and created the Arabian plate. Analysis of the movement of the Arabian plate gives us good insight into different earthquake types.
CONTINENT-CONTINENT COLLISION EARTHQUAKES The Red Sea and Gulf of Aden areas may not have many large earthquakes, but their spreading centers are responsible for shoving the Arabian plate into Eurasia, causing numerous devastating earthquakes there. The rigid continental rocks of the Arabian plate are driven like a wedge into the stiff underbelly of Eurasia. The force of this collision uplifts mountain ranges (e.g., Caucasus and Zagros in figure 4.5 ) and moves many faults that create the killer earthquakes typical of this part of the world.
Black Sea
Mediterranean Sea
Africa
Turkey
ConvergenceTransform fault Left Iateral motion
Collision at
Spreading at trailing edge
Arabian plate Transform
fault
Right lateral motion
D ivergence
40°
30°
30° 40° 50° 60°
20°
10°
leading edge
Iran
Figure 4.21 The Arabian plate pulls away from Africa, pushes into Eurasia, slices through the Holy Land with a transform fault, and squeezes Turkey westward.
Earthquake Fatalities in Iran, 1962–2011
Fatalities Date Location 612 22 Feb 2005 Zarand
41,000 26 Dec 2003 Bam
50,000 21 Jun 1990 Rudbar-Tarom
3,000 28 Jul 1981 Kerman
3,000 11 Jun 1981 Golbas
25,000 16 Sep 1978 Tabas-e-Golshan
5,000 24 Nov 1976 Northwest
5,044 10 Apr 1972 Fars
12,000 31 Aug 1968 Khorasan
12,225 1 Sep 1962 Buyin-Zara
156,881 Total deaths
TABLE 4.3
abb22878_ch04_079-109.indd 94abb22878_ch04_079-109.indd 94 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 95
each year. The rocks along the fault tend to store stress until they can’t hold any more, and then they rupture in an earth- quake-producing fault movement. How often do these earth- quakes occur? Table 4.4 is a partial list from Amos Nur of Stanford University.
Transform-Fault Earthquakes The transform faults forming the sides of some tectonic plates have dominantly horizontal movements that cause major earth- quakes. Examples include the Alpine fault of New Zealand, the San Andreas fault in California, the North Anatolian fault in Turkey, and the Enriquillo-Plantain Garden fault in Haiti.
long as the Red Sea has been opening. During that time, there has been 105 km (65 mi) of offset, and 40 km (25 mi) of this movement has happened in the past 4.5 million years. This computes to an average slip (movement) rate of more than 5 mm/yr over the longer time frame or 9 mm/yr over the more recent time span. However, the rough, frictionally resistant faults do not easily glide along at several millimeters
Gulf of Aqaba
Dead Sea
Sea of Galilee
Damascus
Beirut
Mediterranean Sea
Jerusalem
Antioch
Red Sea Suez
G ulf
of
100 Km0
Some Earthquakes in the Holy Land
Year Magnitude Year Magnitude 1927 6.5 1068 6.6
1834 6.6 1033 7.0
1759 6.5 749 6.7
1546 6.7 658 6.2
1293 6.4 363 7.0
1202 7.2 31 bce 6.3
759 bce 7.3
Source: “When the Walls Came Tumbling Down” (1991). [Video] Amos Nur, Stanford University.
TABLE 4.4
Figure 4.22 Map of the Dead Sea fault zone. Notice that the subparallel faults have pull-apart basins in the steps between faults. The Dead Sea basin is deep; it has a 7 km (greater than 4 mi) thick infill of sediments below its water. On 21 November 1995, a magnitude 7.2 earthquake in the Gulf of Aqaba killed people as far away as Cairo, Egypt.
Figure 4.23 Space shuttle view of the northern Red Sea. The Nile River is in the center left; the Suez Canal, in the middle; and the Gulf of Aqaba, pointing toward the Dead Sea in the upper right. Photo from NASA.
abb22878_ch04_079-109.indd 95abb22878_ch04_079-109.indd 95 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
96 Chapter 4 Plate Tectonics and Earthquakes
were doomed during construction. Building construction was not supervised, allowing bad materials to be used and bad construction practices to be employed. Bad materials include brittle steel; not enough cement; and too much dirty/salty sand in the concrete mixture. Bad construction practices include stopping vertical rods of reinforcement steel at horizontal floors just where they needed to be stron- gest and to be connected. But the major cause of deaths was widespread use of a concrete-blocks-as-filler style of slab construction. Weak, cheap concrete blocks containing large open spaces were placed in the mold before more expensive concrete was poured around them. These blocks just hang in place, supported only by the concrete around them. There are three major problems with these floor slabs (horizontal): (1) they are weak, (2) they are heavy, and (3) they are sup- ported by flimsy concrete columns (vertical) that undergo shear failure when shaken. When one floor collapses, its added weight causes the underlying floor to collapse, and so on.
Tectonic Setting The Caribbean tectonic plate is a small one moving eastward below the huge North American plate (see center of figure 2.14). The northern boundary of the Caribbean plate is split between two parallel, east-west trending faults that pass through Haiti with a total slip of 2 m (6.6 ft) per century. The 2010 seism was caused by left-lateral movement with some compression along the southern of these faults, the Enriquillo- Plantain Garden fault.
TURKEY, 1999: SERIAL EARTHQUAKES A warm and humid evening made sleep difficult, so many people were still up at 3:01 a.m. on 17 August 1999 near the Sea of Marmara in the industrial heartland of Turkey. They were startled by a ball of flame rising out of the sea, a loud explosion, sinking land along the shoreline, and a big wave of water. Another big rupture moved along the North Anatolian fault as a magnitude 7.4 earthquake. This time the fault ruptured the ground surface for 120 km (75 mi), with the south side of the fault moving westward up to 5 m (16.5 ft) ( figure 4.26 ). Several weeks later, after evening prayers for Muslims, another segment of the North Anatolian fault ruptured in a 7.1 magnitude earthquake. The two devas- tating events combined to kill more than 19,000 people and cause an estimated $20 billion in damages.
Why were so many people killed? Bad buildings col- lapsed. Industrial growth in the region attracted hordes of new residents who, in turn, caused a boom in housing construction. Unfortunately, many residential buildings were built on top of soft, shaky ground, and some building contractors cut costs by increasing the percentage of sand in their concrete, causing it to crumble as the ground shook.
HAITI, 2010: EARTHQUAKES DON’T KILL, BUILDINGS DO Year 2010 began with a horrifying event. The earth shook in a 7.0M w event in the Republic of Haiti and much of its capital city of Port-au-Prince collapsed, killing an estimated 230,000 people, seriously injuring another 300,000, and displacing 1.1 million more people. What lies behind this tragedy?
In 1751, most of the buildings in Port-au-Prince were destroyed in an earthquake. In 1770, another earthquake demolished most of the reconstructed city. In response to this double destruction, the French authorities required that buildings be constructed with wood, and they banned the use of construction relying on concrete (masonry).
Haiti gained independence from France in 1804. During its two centuries of independence, the Haitian population grew to 9.35 million people, most of whom suffer poverty, low rates of literacy, and life in poorly constructed, concrete buildings. The lessons of their past about building construction were forgotten.
In 2008, four tropical storms hit Haiti, killing 800 people, displacing 10% of the population, and reducing eco- nomic output by 15%. But for every bad, there is a worse—and life became much worse at 4:53 pm on Tuesday, 12 January 2010, when the powerful earthquake occurred. About 250,000 houses collapsed and another 30,000 commercial buildings fell ( figure 4.24 ). The buildings that collapsed were great and small, rich and poor. Destruction ranged from shacks in shantytowns to the National Palace, the National Assembly building, the United Nations headquarters, the main Catholic cathedral, the upscale Hotel Montana, the Citibank building, schools, hospitals, both fire stations, the main prison, and more.
The death total from this earthquake is more than dou- ble that from any previous 7M earthquake anywhere in the world. The deaths were due to bad buildings, all of which
Figure 4.24 Houses collapsed as poorly made concrete floors, support beams, and walls all failed; Port-au-Prince, Haiti, January 2010. Photo by James L. Harper, Jr., U.S. Air Force.
abb22878_ch04_079-109.indd 96abb22878_ch04_079-109.indd 96 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 97
earthquake affecting Istanbul has a 62(+/−15)% probability of occurring within the next 30 years.
SAN ANDREAS FAULT TECTONICS AND EARTHQUAKES The plate-tectonic history of western North America explains why earthquakes occur. As the Atlantic Ocean basin widens further, both North and South America move westward into the Pacific Ocean basin, helping reduce its size (see figures 2.14 and 4.2). At 30 million years ago, most of the northern portion of the Farallon plate had subducted eastward beneath North America (figure 4.27). At about 28 million years ago, the first segment of the Pacific spreading center collided with North America at about the site of Los Angeles today. The spreading centers to the north and south still operated as before. What con- nected the northern and southern spreading centers? A transform fault, specifically the ancestor of the San Andreas fault.
In the last 5.5 million years, the Gulf of California has opened about 300 km (190 mi). This rifting action has torn Baja California and California west of the San Andreas fault (including San Diego, Los Angeles, and Santa Cruz) from the North American plate and piggybacked them onto the Pacific plate ( figure 4.27 ). The Gulf of California continues to open and is carrying the western Californias on a Pacific plate ride at about 56 mm/yr (2.2 in/yr).
The San Andreas fault is part of a complex system of sub-parallel faults ( figure 4.28 ). The San Andreas fault
The North Anatolian fault is not on the Arabian plate, but it is caused by that plate (see figure 4.21 ). As the Arabian plate pushes farther into Eurasia, Turkey is forced to move westward and slowly rotate counterclockwise in escape tectonics. Bounded by the North Anatolian fault in the north and by the east Anatolian fault in the southeast, Turkey is squeezed westward like a watermelon seed from between your fingers.
The North Anatolian fault is a 1,400 km (870 mi) long fault zone made of numerous subparallel faults that split and combine, bend and straighten. A remarkable series of earthquakes began in 1939 near the eastern end of the fault with the magnitude 7.9 Erzincan earthquake, which killed 30,000 people. Since 1939, 11 earthquakes with magni- tudes greater than 6.7 have occurred as the fault ruptures westward in a semiregular pattern that is unique in the world ( figure 4.26 ). At intervals ranging from 3 months to 32 years, more than 1,000 km (620 mi) of the fault has moved in big jumps.
What is likely to happen next? There is every reason to expect the fault rupture to keep moving to the west, ever closer to Istanbul. The Sea of Marmara fills a basin partly created by movements along subparallel strands of the North Anatolian fault. The region had big earthquakes in 1063, 1509, and 1766. The next big earthquake will likely occur near Istanbul, a city of 13 million people and growing rapidly. In the last 15 centuries, Istanbul has been heavily damaged by 12 earthquakes. Calculations indicate that the next big
Historical Perspective It is interesting to ponder the effects of earthquakes in the Holy Land on the thinking of the religious leaders in this region, which is the birthplace of Judaism and Christianity and an important area to Islam. Imagine the great early leaders living in stiff, mud-block and stone buildings along one of the world’s major strike-slip faults. They under- stood little about the workings of Earth, yet they had to explain and interpret events that destroyed entire cities and killed many thou- sands of people. It is not surprising that many of them interpreted the disastrous events of their times as directly due to "the hand of God."
Let's use today’s understanding of plate tectonics and fault movements to think about past events. For example, how might we interpret the account of Joshua leading the Israelites into the prom- ised land, specifically the famous event when the walls of the oasis city Jericho came tumbling down? Is it possible that during the long siege of Jericho, an earthquake knocked down the walls of the city, killing and disabling many of the residents and allowing Joshua's army to enter and take over? Residents of cities, not troops camped in the surrounding fields, suffer injuries and deaths from the collapse of buildings during an earthquake. Recent historical and archaeo- logical investigations in the Holy Land have shown that many of the destroyed buildings and cities of the past did not meet their ends by time or humans alone; many fell to earthquakes ( figure 4.25 ).
Side Note
Figure 4.25 Building damage in Beit She’an, Israel, caused by the earthquake in 749 CE (common era). Photo by Thomas K. Rockwell.
abb22878_ch04_079-109.indd 97abb22878_ch04_079-109.indd 97 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
98 Chapter 4 Plate Tectonics and Earthquakes
the land offset by the 1906 movements was smoothed out and built upon ( figure 4.30 ). In the vertical plane, the fault move- ment completely ruptured the 15 to 20 km thick brittle layer in the region. The amount of fault movement in 1906 died out to zero at the northern and southern ends of the rupture.
Today, the San Francisco section of the San Andreas fault has a deficit of earthquakes. Apparently this is a “locked” section of the fault (see figure 4.28 ). Virtually all the stress from plate tectonics is stored as elastic strain for many decades until the fault finally can take no more and ruptures in a big event that releases much of its stored energy in a catastrophic movement.
proper is a 1,200 km (750 mi) long, right-lateral fault. In 1906, the northernmost section of the fault broke loose just offshore of the city of San Francisco, rupturing northward and southward simultaneously ( figure 4.29 ). When it stopped shifting, the ground between Cape Mendocino and San Juan Bautista had been ruptured; this is a distance of 400 km (250 mi). The earthquake had a moment magnitude esti- mated at 7.8 resulting from 110 seconds of fault movement. When movement stopped, the western side had shifted north- ward a maximum of 6 m (20 ft) horizontally. In the peninsula south of San Francisco, fault movements have formed elon- gate topographic low areas now filled by lakes, and some of
D ea
th s
200,000
10,000
20,000 30,000
2,000 4,000 6,000 8,000
1999 Aug. Nov.
1967 1957
1944 1951
1943 1942
1939
1992
Istanbul Izmit
0 100 km
7.4 7.1
7.1
7.0
7.3
7.3 7.1
7.9
6.9
Ankara
Nort h Ana
tolian fault
TURKEY
Sea of Marmara
Black Sea
Figure 4.26 The North Anatolian fault accommodates the movement of Turkey westward into the Mediterranean basin. Note the time sequence of the fault ruptures from east to west. What does the near future hold for Istanbul? Photo (left) © John A. Rizzo/Getty Images RF and photo (center and right) by Roger Bilham/NOAA.
abb22878_ch04_079-109.indd 98abb22878_ch04_079-109.indd 98 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 99
The San Andreas fault segment north of Los Angeles is another locked zone that is deficient in earthquake activity ( figure 4.28 ). However, on 9 January 1857, this segment of the fault broke loose at its northwestern end, and the rupture propagated southeastward in the great Fort Tejon earthquake with a magnitude of about 7.9. Due to the one-way advance of the rupture front, the fault movement lasted almost 3 minutes. The ground surface was broken for at least 360 km (225 mi), and the maximum offsets in the Carrizo Plain ( figure 4.31 ) were a staggering 9.5 m (31 ft). One of the offset features was a circular corral for livestock that was split and shifted to an
The San Andreas fault has different behaviors along its length. The section to the south of San Francisco ( figure 4.28 ) has frequent small-to moderate-size earthquakes. This is a “creeping” section of the fault where numerous earthquakes accommodate the plate-tectonic forces before they build to high levels. The creeping movements of the fault are shown by the millimeters per year of ongoing offset of sidewalks, fences, buildings, and other features. Earthquakes in this fault segment do not seem to exceed magnitude 6. These are still significant seisms, but they are small compared to events on adjoining sections of the fault.
N or
th A
m er
ic an
p la
te
N or
th A
m er
ic an
p la
te
F ar
al lo
n pl
at e
N or
th A
m er
ic an
p la
te
N or
th A
m er
ic an
p la
te
F ar
al lo
n pl
at e
F ar
al lo
n pl
at e
B aj
a C
al ifo
rn ia
20 million years ago
10 million years ago
Present 30 million years ago
Pacific plate
Pacific plate
Pacific plate
Pacific plate
Pacific plate
Los Angeles San Diego
Los Angeles
San Francisco
Santa Cruz
Juan de Fuca plate
Ju an
d e
Fu ca
pl at
eJu an
d e
Fu ca
pl at
e
Rivera plate
Cocos plate
Los Angeles
Trench
Los Angeles
Trench Trench
Trench
T re
nc h
T re
nc h
T re
nc h
M
M M
R R
R
Figure 4.27 Collision of the Pacific Ocean basin spreading center with the North American plate: (a) 30 million years ago, the first spreading-center segment nears Southern California; (b) 20 million years ago, a growing transform fault connects the remaining spreading centers; (c) 10 million years ago, the Mendocino (M) and Rivera (R) triple junctions continue to migrate north and south respectively; (d) at present, the long transform fault is known as the San Andreas fault. (Interpretations based on the work of Tanya Atwater.) Source: Kious, W. J., and Tilling, R. I., This Dynamic Earth. US Geological Survey, p. 77.
abb22878_ch04_079-109.indd 99abb22878_ch04_079-109.indd 99 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
100 Chapter 4 Plate Tectonics and Earthquakes
California
Locked 1906
San Francisco
C reep
San Andreas
1857fault Locked
Garl ock
Jacinto
San
fault San Diego
Los Angeles
O w
ens V
alley
fault
fault
195 2
Wh ite
Wo lf
1872
Pacific Ocean
Segments on which slip occurred during great earthquakes of 1857, 1872, and 1906
Key:
1992
Figure 4.28 Historic behavior of some California faults. The northern “locked” section of the San Andreas fault ruptured for 250 mi in 1906 (magnitude 7.8). The central “creeping” section has frequent smaller earthquakes. The south-central “locked” section ruptured for 225 mi in 1857 (magnitude 7.9). The southernmost San Andreas awaits a major earthquake. The Owens Valley fault ruptured for 70 mi in 1872 (magnitude 7.3). A magnitude 7.5 seism occurred on White Wolf fault in 1952, and a magnitude 7.3 seism happened in the Mojave Desert in 1992. Source: “The San Andreas Fault,” US Geological Survey.
Figure 4.29 Looking south-southeast down the San Andreas fault. View is over Bodega Head and Tomales Bay toward the epicenter of the 1906 San Francisco earthquake. Photo by John S. Shelton.
Figure 4.30 Looking southeast along the trace of the San Andreas fault. The San Francisco airport and part of San Francisco Bay are in left center. Linear lakes in right center (e.g., Crystal Springs Reservoir) are in the fault zone. In bottom center, the land offset by the 1906 fault movement was bulldozed and covered with houses! Photo by John S. Shelton.
Figure 4.31 The San Andreas fault slashes across the Carrizo Plain. Notice the ridges and basins caused by local squeezing and pulling apart. Photo courtesy of Pat Abbott.
abb22878_ch04_079-109.indd 100abb22878_ch04_079-109.indd 100 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 101
bend tend to be infrequent and large. This left step in the San Andreas zone also causes the fault plane to be inclined 70° to the southwest ( figure 4.33 ). The fault movement began at 18.5 km (11.5 mi) depth and slipped for 2.3 m (7.5 ft). The motion can be resolved into 1.9 m (6.2 ft) of horizontal movement (strike slip) and 1.3 m (4.3 ft) of vertical movement (reverse slip). Stated dif- ferently, the western or Pacific plate side moved 6.2 ft to the northwest, and a portion of the Santa Cruz Mountains was uplifted 36 cm (14 in). Although the fault did not rupture the surface, the uplifted area was 5 km (3 mi) wide and had numerous fractures in the uplifted and stretched zone. Many of the cracked areas became the sites of landslides.
The mainshock had a surface-wave magnitude (M s ) of 7.1 and a moment magnitude (M w ) of 6.9; numerous aftershocks followed, as is typical for large earthquakes. The Loma Prieta
S-shape by the fault movement. In 1857, the region was sparsely settled, so the death and damage totals were small. The next time a great earthquake occurs here, the effects may be disastrous.
The southernmost segment of the San Andreas fault, from San Bernardino to its southern end at the Salton Sea, has not generated a truly large earthquake in California’s recorded history. But we can extend our knowledge of earth- quakes into the prehistoric past by measuring offsets in sedi- mentary rock layers. For example, we’ve learned that the last truly big earthquake on the southern San Andreas fault occurred about the year 1690. These techniques will be discussed in chapter 5.
World Series (Loma Prieta) Earthquake, 1989 In 1989, the World Series of baseball was a Bay Area affair. It pitted the American League champion Oakland Athletics against the National League champion San Francisco Giants. Game 3 was scheduled in San Francisco’s Candlestick Park, where the Giants hoped the home field advantage would help them win their first game. It was Tuesday, 17 October, and both teams had finished batting practice, which was watched by 60,000 fans at the park, along with a television crowd of another 60 million fans in the United States and millions more around the world. At 5:04 p.m., 21 minutes before the game was scheduled to start, a distant rumble was heard, and a soft thunder rolled in from the southwest, shaking up the fans and stopping the game from being played. San Francisco was experiencing another big earthquake, and this time, it shared it with television viewers. After the earthquake, the San Franciscans at Candlestick Park broke into a cheer, while many out-of-staters were seen heading for home.
What caused this earthquake? An 83-year-long pushing match between the Pacific and North American plates resulted in a 42 km (26 mi) long rupture within the San Andreas fault system. The southernmost section of the fault zone that moved in 1906 had broken free and moved again. There were several different aspects to the 1989 earthquake: (1) The fault rupture took place at depth; (2) the fault movement did not offset the ground surface; (3) there was significant vertical movement; and (4) the fault rupturing lasted only 7 seconds, an unusually short time for a magnitude 6.9 event.
Movement occurred in a gently left-stepping constrain- ing bend of the San Andreas fault zone ( figure 4.32 ). Long- term compressive pressures along this left step have uplifted the Santa Cruz Mountains. This step in the San Andreas fault is near where the Calaveras and Hayward faults split off and run up the east side of San Francisco Bay. The epicenter of the 1989 seism was near Loma Prieta, the highest peak in the Santa Cruz Mountains. Loma Prieta is the official name of this earthquake; it follows the rule of taking the name from the most prominent geographic feature near the epicenter. Nonetheless, this event remains known to many people as the World Series earthquake.
It is difficult for a fault to move around a left-stepping bend. Constraining bends commonly “lock up”; thus, movements at a
**
Santa Rosa
Marina District
H ealdsburg-R
odgers C reek fault
Sacramento
G reen V
alley fault
Pacific Ocean
0
0 30 km
20 mi
I-880 structure
San Francisco
Oakland
A ntioch
fault
C oncord
fault Hayw
ard fault
San Jose
Los Gatos Loma Prieta
1157m
San Andreas fault
Zayante fault
Epicenter
Santa Cruz Watsonville
C alaveras fault
S an G
regorio fault Segment that slipped in 1989 Loma Prieta
earthquake
Sargent fault
123°
38°
37°
122°
Figure 4.32 Map showing the epicenter of the World Series (Loma Prieta) earthquake. The San Andreas fault takes an 8° to 10° left step in the ruptured section. The left step also is where the Calaveras and Hayward faults split off from the main San Andreas trend. Source: “Lessons Learned from the Loma Prieta Earthquake of October 17, 1989” in US Geological Survey Circular 1045, 1989.
abb22878_ch04_079-109.indd 101abb22878_ch04_079-109.indd 101 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
102 Chapter 4 Plate Tectonics and Earthquakes
south moves frequently, generating numerous small earth- quakes. But the same plate-tectonic stresses affecting the creep zone also affect the locked or seismic-gap zone. How does a locked zone catch up with a creep zone? By infrequent but
area had been a relatively quiet zone for earthquakes since the 1906 fault movement ( figure 4.34 ); before 1989, the Loma Prieta region had been a seismic gap. As the numerous epicen- ters in figure 4.34 a show, the San Andreas fault section to the
Direction of plate
movement
Surface cracks showing direction
of movement San Andreas fault
Fault slip at depth in
this earthquake
Hypocenter
Fault rupture
Pac ific
plat e
Nor th
11.5
mile s
3.7 mile
sUplif t =
14 in ches
Ame rica
n
plat e
Form er ground surface
7.5 ft total slip
6.2 ft
4.3 ft
N
San FranciscoNorth Loma Prieta Parkfield South
0 10 20
0 10 20
D ep
th (
km )
D ep
th (
km )
0 50 150 250 350 Distance (km)
(a)
(b)
Figure 4.33 Schematic diagram of fault movement within the San Andreas zone in the World Series earthquake. The San Andreas fault dips 70° southwest because of the left-step bend. Fault movement began at 18 km (11.5 mi) depth and moved 1.9 m (6.2 ft) horizontally and 1.3 m (4.3 ft) vertically. Fault movement died out upward and did not rupture the ground, although the surface bulged upward 36 cm (14 in). Think three-dimensionally here: because of the dipping fault plane, will the epicenter plot on the ground-surface trace of the San Andreas fault? No. Source: “Lessons Learned from the Loma Prieta Earthquake of October 17, 1989” in US Geological Survey Circular 1045, 1989.
Figure 4.34 Cross-sections of seismicity along the San Andreas fault, 1969 to early 1989. (a) Notice the dense concentrations of hypocenters in the central creeping section of the fault from south of Loma Prieta to Parkfield, as well as the “seismic gap” in the Loma Prieta area. (b) Notice the deep hypocenter (in red) of the 1989 mainshock plus the numerous aftershocks. Putting the two cross-sections together fills the seismic gap. Are there other seismic gaps in cross-section (a)? Yes, south of San Francisco in the Crystal Springs Reservoir area (see figure 4.30 ), just west of the densely populated midpeninsula area. When will this seismic gap be filled? Source: “Lessons Learned from the Loma Prieta Earthquake of October 17, 1989” in US Geological Survey Circular 1045, 1989.
abb22878_ch04_079-109.indd 102abb22878_ch04_079-109.indd 102 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 103
large fault movements. Notice in figure 4.34 b how the World Series (Loma Prieta) mainshock and aftershocks filled in the seismic gap in cross-section (a). This demonstrates some merit for the seismic-gap method as a forecasting tool. Figure 4.34 also shows another seismic gap, south of San Francisco in the heavily populated midpeninsula area (this is the area of elon- gate lakes shown in figure 4.30 ). The 1989 fault movement has increased the odds by another 10% for a large earthquake in the Crystal Springs Reservoir area in the next 30 years.
In the World Series earthquake, the fault ruptured at greater than 2 km/sec in all directions simultaneously, upward for 13 km (8 mi), and both northward and southward for over 20 km (13 mi) each. Table 3.7 indicates that earthquakes with magnitudes of 7 usually rupture for about 20 seconds; this radially spreading, 6.9-magnitude rupture lasted only 7 seconds. Had it lasted the expected 20 seconds, numerous other large buildings and the double-decker Embarcadero Freeway in San Francisco would have failed catastrophically. As it was, the event left 67 people dead or dying, 3,757 injured, and more than 12,000 homeless; caused numerous landslides; disrupted transportation, utilities, and communications; and caused about $6 billion in damages.
Building Damages In the epicentral region, serious damage was dealt to many older buildings. The short-period P and S waves wreaked their full effects on low buildings built of rigid materials. Common reasons for failure included poor connections of houses to their foundations, buildings made of unreinforced masonry (URM) or brick-facade construction, and two-to-five-story buildings deficient in shear-bearing internal walls and sup- ports. In Santa Cruz, four people died, and the Pacific Garden Mall, the old city center of historic brick and stone buildings that had been preserved and transformed into a tourist mecca, was virtually destroyed.
From the epicentral region, the seismic waves raced outward at more than 3 mi/sec. Some longer-period shear waves remained potent even after traveling 100 km (more than 60 mi). Upon reaching the soft muds and artificial-fill foundations around San Francisco Bay, these seismic waves had their vibrations amplified. Ground motion at some of these soft-sediment foundation sites was 10 times stronger than at nearby sites on rock.
Marina District The Marina District is one of the most beautiful areas in San Francisco. It sits on the northern shore of the city next to parks, the Golden Gate Bridge, and the bay itself. In this desirable and expensive district, five residents died, building collapses were extensive, and numerous building-eating fires broke out due to: (1) amplified shaking, (2) deformation and liquefaction of artificial-fill foundations, and (3) soft first-story construction, which led to building collapses (figure 4.35).
Much of the Marina District is built on artificial fill dumped onto the wetlands of the bay to create more land
Magnified Magnified
(a)
Figure 4.35 (a) Water-saturated sediment usually rests quietly (left). However, when seismic waves shake, sand grains and water can form a slurry and flow as a liquid (right). When earth materials liquefy, building foundations may split and buildings may fail. (b) A typical Marina District building collapse. Three residential stories sat above a soft first story used for car parking; now, the four-story building is three stories tall. Photo from Dames and Moore.
abb22878_ch04_079-109.indd 103abb22878_ch04_079-109.indd 103 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
104 Chapter 4 Plate Tectonics and Earthquakes
for development. Ironically, much of the artificial fill was the debris from the San Francisco buildings ruined by the 1906 earthquake. Seismic waves in 1989 were amplified in this artificial fill. Some fill underwent permanent defor- mation and settling, and some formed slurries as under- ground water and loose sediment flowed as fluids in the process of liquefaction ( figure 4.35 a). Liquefaction in the Marina District in 1989 brought to the surface pieces of glass, tar paper, redwood, and other debris from 1906 San Francisco.
The central cause of building failure was flawed design. Because the Marina District is home to many affluent people, they need places to park their cars. But where? The streets are already overcrowded, and basement parking garages would be below sea level and thus flooded. A common solution has been to clear obstructions from the first stories of buildings to make space for car parking. That means removing the internal walls,
Figure 4.36 The Cypress double-decker section of Interstate 880 in Oakland was completed in 1957. It failed in the 1989 seism and dropped 1.25 mi of upper roadbed onto the lower roadbed, crushing many vehicles and people. Photo from Dames and Moore.
lateral supports, and bracing needed to support the upper one to four sto- ries. This creates a “soft” first story, so that in an earthquake, buildings simply pancake and become one story shorter ( figure 4.35 b). It is estimated that there are 2,800 blocks of soft first-story residences in San Francisco today and another 1,500 blocks in Oakland.
Interstate 880 The most stunning tragedy associated with the World Series earthquake was the crushing of 42 people during the collapse of a double-decker por- tion of Interstate 880 in Oakland
Figure 4.38 The support columns of the Interstate 880 structure failed at the joints. There were 20 #18 bars of steel in each column, but they were discontinuous at the joints and failed there. Photo from Dames and Moore.
10
880
580
80
To San Francisco
Cypress structure
Oakland
80
Seconds
0
Bedrock
Soft mud
Sand and gravel
10
0
0 1.6 km
1 mi
Figure 4.37 The portion of Interstate 880 elevated roadway built on top of soft bay mud collapsed (dashed purple line), while the portion resting on sand and gravel still stood (solid purple line). Notice how the shaking was amplified in the soft mud.
abb22878_ch04_079-109.indd 104abb22878_ch04_079-109.indd 104 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 105
( figure 4.36 ). The elevated roadway was designed in 1951 and completed in 1957. A 2 km (1.25 mi) long section collapsed: 44 slabs of concrete roadbed, each weighing 600 tons, fell onto the lower roadbed and crushed some vehicles to less than 30 cm (1 ft) high. The section that collapsed was built on young, soft San Francisco Bay mud. The elevated freeway structure had a natural resonance of two to four cycles per second; the bay-mud foundation produced a five to eight-fold amplification of shaking in that range. The seismic waves excited the mud ( figure 4.37 ), causing the heavy structure to sway sharply. The portion of I-880 elevated roadway built on firmer sand and gravel stood intact; the portion standing on soft mud collapsed catastrophically.
The weak foundation was compounded by a flawed struc- tural design. The joints where roadbeds were connected to concrete support columns were not reinforced properly. Cracks initiated at the joints caused failure of supporting columns, which slid off the crushed areas of the joints and dropped the upper roadbed onto the lower level ( figure 4.38 ). Was this bridge failure a surprise? Not really. The lessons had been learned 18 years earlier in the 1971 San Fernando earthquake, but no one had corrected this disaster-in-waiting.
An ironic and deadly footnote to this disaster lay in the mode of failure. There was a delay between the initial shock
The San Francisco Earthquake of 1906 Early in the 20th century, San Francisco was home to about 400,000 people who enjoyed a cosmopolitan city that had grown during the economic boom times of the late 19th century. During the evening of 17 April 1906, many thrilled to the special appearance of Enrico Caruso, the world’s greatest tenor, singing with the Metropolitan Opera Com- pany in Bizet’s Carmen. But several hours later, at 5:12 a.m., the initial shock waves of a mammoth earthquake arrived to begin the destruction of the city. One early riser told of see- ing the earthquake approach as the street before him literally rose and fell like a series of ocean swells moving toward shore.
During a noisy minute, the violently pitching Earth emitted dull booming sounds joined by the crash of human- made structures. When the ground finally quieted, people went outside and gazed through a great cloud of dust to view the destruction. Unreinforced masonry buildings lay col- lapsed in heaps, but steel-frame buildings and wooden struc- tures fared much better. Another factor in the building failures was the nature of the ground they were built on. Destruction was immense in those parts of the city that were built on artificial fill that had been dumped onto former bay wetlands or into stream-carved ravines.
As repeated aftershocks startled and frightened the survivors, another great danger began to grow. Smoke arose from many sites as fires fed on the wood-filled rubble. Unfortunately, the same earthquake waves that wracked the buildings also broke most of the water lines, thus hindering attempts to stop the growing fires. From the business dis- trict and near the waterfront, fires began their relentless intrusion into the rest of the city. Desperate people tried dynamiting buildings to stop the fire’s spread, but they only provided more rubble to feed the flames or even blew
flaming debris as far as a block away, where it started more fires.
The fires did about 10 times as much damage as the earthquake itself; fire destroyed buildings covering 490 city blocks. More than half the population lost their homes. Death and destruction were concentrated in San Francisco, where 315 people died, but the affected area was much larger. About 700 deaths occurred in a 430 km (265 mi) long belt of land running near the San Andreas fault. Towns within the high-intensity zone, such as San Jose and Santa Rosa, were heavily damaged, yet other cities to the east of the narrow zone, such as Berkeley and Sacramento, were spared signifi- cant damage. Problems continued in the months that fol- lowed as epidemics of filth-borne diseases sickened Californians; more than 150 cases of bubonic plague were reported. When all the fatalities from earthquake injuries and disease are included, the death total from the earthquake may have been as high as 5,000.
Total financial losses in the event were almost 2% of the US gross national product in 1906; for comparison, Hurricane Katrina economic losses were much less than 1%. Politicians and the press in their desire to restore the city called the disaster a fire-related event and listed the death total at about 10% of actual life loss. In their desire to rebuild, the emphasis was on quickness, not on increasing safety. This problem haunts us today because much of the early rebuilding was done badly and is likely to fail in the next big earthquake.
One of the intriguing aspects of disasters is their ener- gizing effects on many survivors. Hard times shared with others bring out the best in many people. Shortly after this earthquake, the resilient San Franciscans were planning the Panama-Pacific International Exposition that was to impress the world and leave behind many of the beautiful buildings that tourists flock to see today. You can’t keep a good city down.
A Classic Disaster
abb22878_ch04_079-109.indd 105abb22878_ch04_079-109.indd 105 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
106 Chapter 4 Plate Tectonics and Earthquakes
earthquakes, and it is quite different from the 20th-century record. During the 19th century, earthquakes with magni- tudes greater than 6 were much more common ( figure 4.39 ). There were seven destructive seisms in the 70 years before the 1906 San Francisco earthquake, averaging a large earth- quake every decade. Then came the monstrous movement of the San Andreas fault in 1906. This 250-mile-long rupture removed so much of the plate-tectonic stress stored in the rocks that several decades of the 20th century were effec- tively free of large earthquakes ( figure 4.40 ). But large earth- quakes returned to the southern part of the Bay Area beginning in the 1970s ( figure 4.39 ). We can identify three patterns in these data.
Pattern 1 Common Large Earthquakes versus Rare Giant Shakes The movement of the Pacific plate past the North American plate in the Bay Area seems to be satisfied by either a magnitude
and the final collapse, which allowed some people a brief time to plan. Some maneuvered their vehicles under beams next to support columns, and others got out of their cars and walked under the same supports, thinking that these were the strongest parts of the structure; but the steel bars in the sup- port columns were discontinuous. Tragically, these were the weak spots, where failure was most catastrophic, and no one survived there.
BAY AREA EARTHQUAKES— PAST AND FUTURE The historic record of California earthquakes is accurate only back to about 1850, and thus is shorter than the recurrence times for major movements on most faults. Nonetheless, the San Francisco Bay Area has enough information contained in newspaper accounts, diaries, personal letters, and similar sources to piece together a fairly accurate history of 19th-century
A ntioch
fault
fault
fault
fault C
oncord fault
H ayward
San Andreas fault
S an
G regorio
fa u lt
Zayante fault
C alaveras fault
6.4M
1868
6.5M
Santa Cruz
1989 7.0M
1865
San Jose
1911 6.5M
1836 6.5M
1984 6.2M
1979 5.9M
1861 5.8M
1858 6.1M
6.9M
Hayward
Oakland
San Francisco
1838 6.8M
1892 Sacramento
122°123°
38°
37°
Pacific Ocean
Santa Rosa
1906 7.8M
Sargent fault
H ealdsburg - Rodgers C
reek
G reen Valley
0
0 20 km
20 mi
Figure 4.39 Locations and approximate sizes of some larger Bay Area earthquakes. Source: US Geological Survey.
abb22878_ch04_079-109.indd 106abb22878_ch04_079-109.indd 106 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Transform-Fault Earthquakes 107
Figure 4.40 Distribution of earthquakes with magnitudes greater than 5.5 near San Francisco Bay, 1849–2011. The index map shows the area of earthquake epicenters.
M ag
ni tu
de
123
1850 1870 1890 1910 1930 1950 1970 1990 2011 5
6
7
8
° 122° 121°
Santa Rosa Vacaville
Livermore
Hollister
38°
37°
0 40 km
Year
San A
ndreas Fault
G reenville
R odgers
Creek Fault
Mount Diablo
Thrust Fault
C o
n co
rd -G
reen
H ayw
ard Fault
Sacramento
San Francisco Bay Region Earthquake Probability
62%Napa
Walnut Creek
Danville
Sonoma
Santa Rosa
Petaluma
Novato
San Rafael
Vallejo
Antioch
Hayward
Palo Alto
San Mateo
OaklandSan Francisco
Pacifica
N
20 km0
20 mi
>10% 4-10%
1-4% <1%
0
Half Moon Bay
Santa Cruz Watsonville
Monterey Bay
Monterey Salinas
Gilroy Extent of rupture
in Loma Prieta Quake
San Jose
3%3%
3%3%
11%11%
10%10%
4%4%
21%21%
Stockton
Tracy
V alley Fault
Fault
C alaveras
S an G
regorio Fault
Fault
S an Francisco Bay
Pacific O cean
Probability for one or more magnitude 6.7 or greater earthquakes from 2003 to 2032. This result incorporates 14% odds of quakes not on shown faults.
Probability in a 30-year period from 2003 to 2032
Explanation
27%27%
Figure 4.41 Probabilities of one or more magnitude 6.7 or larger earthquakes in the San Francisco Bay region, 2003–2032. Source: Redrawn from Working Group on California Earthquake Probabilities, 2003.
abb22878_ch04_079-109.indd 107abb22878_ch04_079-109.indd 107 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
108 Chapter 4 Plate Tectonics and Earthquakes
Summary Most earthquakes are caused by fault movements associated with tectonic plates. Plates have three types of moving edges: (1) divergent at spreading centers, (2) slide-past at transform faults, and (3) convergent at collision zones. The tensional (pull-apart) movements at spreading centers do not produce very large earthquakes. The dominantly horizontal (slide- past) movements at transform faults produce large earth- quakes. The compressional movements at subduction zones and continent-continent collisions generate the largest tec- tonic earthquakes, and they affect the widest areas.
Subduction zones produce the largest number of great earthquakes. In 1923, a subduction movement of the Pacific plate destroyed nearly all of Tokyo and Yokohama; much of the devastation was caused by fires unleashed during build- ing collapses. The largest earthquakes along western North America are due to subduction beneath the continent. The magnitude 9.2 Alaska earthquake in 1964 and Japan earth- quake in 2011 were due to subduction of the Pacific plate, and the magnitude 8.1 Mexico City event in 1985 was caused by subduction of the Cocos plate. The plates subducting beneath Oregon, Washington, and British Columbia gener- ated a magnitude 9 earthquake on 26 January 1700 and will do so again in the future. In 2004, the Sumatra, Indonesia, earthquake and tsunami killed more than 245,000 people.
Continent-continent collisions produce great earthquakes throughout Asia and Asia Minor. The 2005 Kashmir, Pakistan, earthquake killed 88,000; the 2008 Sichuan, China, earth- quake killed 87,500; and the 2001 Gujarat, India, quake killed 20,000. The earthquakes did not kill directly; it was the col- lapse of human-built structures that was deadly. The deadliest earthquake in history occurred in 1556 in Shaanxi Province, China, when the loose, silty sediment into which cave homes had been dug collapsed and flowed, killing 830,000 people.
Deaths from earthquakes are mostly due to building failures. For example, for thousands of years, humans have built stone and mud-block houses along the Dead Sea fault zone (a major transform fault), and for thousands of years, these rigid houses have collapsed during earthquakes, caus- ing many deaths. These geologic disasters have affected the teachings of Judaism, Christianity, and Islam.
The frequent earthquakes of western North America are mostly due to plate tectonics. The westerly moving North American plate has overrun the Pacific Ocean spreading center along most of California. To the south, ongoing spread- ing has torn Baja California from mainland Mexico, and Baja California, San Diego, Los Angeles, and Santa Cruz are now riding on the Pacific plate toward Alaska at 5.6 cm/yr. To the north, spreading still occurs offshore from northernmost
covered by schools, hospitals, city halls, houses, and the University of California. If a seism like the 1868 earthquake occurs soon, the California Division of Mines and Geology estimates that up to 7,000 people might die. The number of deaths will depend in part on the time of day the earthquake occurs. When is the worst time for an earthquake? During the middle of the work and school day, when the maximum number of people are occupying the larger, older structures. When is the best time for an earthquake? During the night, when most people are home and asleep in their beds. In general, California houses handle earthquake shaking quite well because their wood frameworks are flexible and well tied together with nails, bolts, and braces.
Bay Area Earth quake Probabilities For the combined San Francisco Bay region and its 6.8 million people, there is a 62% (+/−10%) chance that a 6.7 magnitude (Northridge-size) earthquake will occur on a fault crossing through the urban area before 2032 ( figure 4.41 ).
The probability of the Hayward fault causing a magni- tude 6.7 or greater earthquake before 2032 is estimated at 27%. The Hayward fault is expected to rupture for about 22 seconds with about 2 m (6 ft) of slip extending down about 13 km (8 mi). The next movement of the Hayward fault will cause tens of billions of dollars in property losses, and deaths may total in the thousands.
6 to 7 earthquake roughly every decade (19th century) or a magnitude 8 earthquake every century (20th century). Which pattern is preferable for this heavily developed and populated region (not that we have any choice)? Which pattern causes the least amount of death, damage, and psychological distress? Will the 21st century be like the 19th or the 20th?
Pattern 2 Pairings of Earthquakes In 1836, the Monterey Bay area experienced a quake of about magnitude 6.5; this was followed two years later on the San Francisco Peninsula with a seism of about magnitude 6.8 (see figure 4.39 ). In the southern Bay Area, a large shallow earthquake near Santa Cruz in 1865 was followed three years later by a shake of about magnitude 6.9 near Hayward. Will this pattern of paired earthquakes reoccur?
Pattern 3 Northward Progression of Earthquakes The large earthquakes of 1865 and 1868 were preceded by five moderate earthquakes that moved northward up the Calaveras fault. Figure 4.39 shows moderate to large earth- quakes that have moved from south to north up the Calaveras fault. Does this repeat pattern suggest an upcoming large seism on the Hayward fault? This region today is populated by more than 2 million people in 10 cities. The fault is
abb22878_ch04_079-109.indd 108abb22878_ch04_079-109.indd 108 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
8. If present seafloor spreading trends continue, what will happen to Baja California, San Diego, Los Angeles, and Santa Cruz?
9. Which part of the United States sits in an opening ocean basin? Evaluate the earthquake threat there.
10. How long were the surface ruptures in the 1906 San Francisco and 1857 Fort Tejon earthquakes? What was the maximum offset of the surface during each quake?
11. Evaluate the earthquake hazards in locked versus creeping segments of the San Andreas fault. Are the biggest cities in locked or creeping segments?
12. Evaluate the seismic gap in the San Andreas fault south of San Francisco.
13. What factors combined to cause the resonance in Mexico City that was so deadly in the 1985 earthquake? How far was the city from the epicenter?
14. Sketch a Marina District (San Francisco) dwelling and explain why so many failed during the 1989 Loma Prieta earthquake.
15. What is usually the worst time of day for a big earthquake to strike a city in the western United States?
16. What are the four stages of formation of an ocean basin? 17. Can one large earthquake trigger others? What is the recent
experience in Indonesia? 18. In the 2011 Japan earthquake, how large an area of plate
moved? What was the maximum slip? What was the earthquake magnitude? When did the last earthquake of this size occur in the same area?
19. What is the largest earthquake measured (see Chile)? 20. Why do so many mega-killer earthquakes occur in the China,
India, Pakistan region? 21. The 2010 Haiti and 1989 Loma Prieta (World Series)
earthquakes were both 7M events. Why were 3,000 times more people killed in the Haiti earthquake?
22. Can we recognize that an earthquake is a foreshock before the mainshock occurs?
23. Sketch a sequence of cross-sections that shows how a continent is split, then separated to form an ocean basin.
Questions for Further Thought 1. How might people with no geologic knowledge, living in stone
houses next to a major fault, explain a disastrous earthquake? 2. How might you use food to create a plate-tectonics model in
your kitchen? 3. Which U.S. states are on the Pacific plate? 4. Which would be the better of two bad choices for an urban
area: a magnitude 6.5 to 7 earthquake every 15 years or a magnitude 8 every century?
5. Why is the zone of active faults so much wider in southern than in northern California?
6. If a magnitude 9 earthquake occurred in the Cascadia subduction zone offshore from the Pacific Northwest, what might happen in Vancouver, Seattle, Portland, and other onshore sites?
7. On 19 September 1985, Mexico City was rocked by a magnitude 8.1 earthquake. Two days later, the city was shaken by a magnitude 7.5 earthquake. Would you consider this a mainshock and an aftershock or twin earthquakes?
8. Is East Africa likely to pull away from the rest of Africa to form a Somali plate?
California, Oregon, Washington, and southern British Columbia. Two separated spreading centers are connected by a long transform fault—the San Andreas fault. Its earthquakes include a magnitude 7.9 caused by a 225 mi long rupture in central California in 1857, a magnitude 7.8 due to a 260 mi long rupture passing through the San Francisco Bay region in 1906, and a magnitude 6.9 unleashed by a 25 mi long rupture near Santa Cruz in 1989.
Major losses of life and property damage are commonly due to problems with buildings. In San Francisco in 1906, unreinforced-masonry buildings collapsed, especially those built on artificial-fill foundations. The worst damage was done by two and one-half days of fires that raged unchecked because ground shaking had broken water pipes, rendering firefighters largely helpless. In Mexico City in 1985, 1- to 2-second-period shear waves caused shaking of 6- to 16-story buildings at the same 1- to 2-second frequency, and shaking was amplified in muddy, former lake-bottom sediments. The resonance of seismic waves and tall build- ings, amplified by soft sediment foundations, caused numer- ous catastrophic failures.
Earthquake numbers and sizes have varied in the San Francisco Bay region. In the 19th century, magnitude 6.5 to 7 events occurred at an average of one per decade; the 20th century was dominated by the magnitude 7.8 event in 1906.
Southern California may have several large earthquakes in the 21st century. The southern segment of the San Andreas fault is the only one not to have a long rupture in historic time. In prehistory, it has ruptured every 250 years on aver- age, but the last big movement was in 1690.
Questions for Further Thought 109
Questions for Review 1. Draw a map of an idealized tectonic plate and explain the
earthquake hazards along each type of plate edge. 2. Sketch a map of the Arabian plate and explain the origin of
the Iranian, Holy Land, and Turkey earthquakes. 3. Explain why earthquakes at subduction zones are many times
more powerful than spreading-center earthquakes. 4. Explain the seismic-gap method of forecasting earthquakes. 5. Which tectonic-plate edges fail most commonly in shear? In
tension? In compression? 6. Why are fires in cities so commonly associated with major
earthquakes? 7. Sketch a plate-tectonic map along western North America
from Alaska through Mexico. Label the spreading centers, subduction zones, and transform faults. Label the maximum earthquakes expected along the coastal zones.
liquefaction 104 rift 82 seismic-gap method 86 slip 86 slurry 104 triple junction 82
Terms to Remember ce 97 cohesion 93 earthquake cluster 87 escape tectonics 97 fault 80 global positioning system (GPS) 86
abb22878_ch04_079-109.indd 109abb22878_ch04_079-109.indd 109 13/12/12 8:23 PM13/12/12 8:23 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Earthquakes Throughout the United States and Canada 5
CHAPTER
Eventually, everything east of the San Andreas fault will break off and fall into the Atlantic Ocean.
—Michael Grant, 1982, San Diego Union
Highway 287 in Montana was destroyed by the Hebgen Lake earthquake on 17 August 1959. Photo from US Geological Survey
LEARNING OUTCOMES Earthquakes occur in many places not related to plate tectonics. After studying this chapter, you should:
• recognize the complexities of fault movements.
• realize the lack of connection between earthquakes and weather.
• know how the dates and magnitudes of prehistoric earthquakes can be determined.
• understand our inability to make short-term predictions of earthquakes.
• know the ways that humans trigger earthquakes.
• realize that earthquakes occur in every state.
• be familiar with the relationship between volcanism and earthquakes.
OUTLINE • How Faults Work
• Thrust-Fault Earthquakes
• Normal-Fault Earthquakes
• Neotectonics and Paleoseismology
• Earthquake Prediction
• Human-Triggered Earthquakes
• Earthquake-Shaking Maps
• California Earthquake Scenario
• Earthquakes in the United States and Canada
• Western North America: Plate Boundary–Zone Earthquakes
• Intraplate Earthquakes: “Stable” Central United States
• Intraplate Earthquakes: Eastern North America
• Earthquakes and Volcanism in Hawaii
In te
r n
al E
n er
g y
abb22878_ch05_110-142.indd 110abb22878_ch05_110-142.indd 110 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
How Faults Work 111
At 1:51 p.m. on Tuesday, 23 August 2011, a 5.8M w earthquake near the town of Mineral in Virginia shook eastern North America and rattled the nerves of millions of people. Office workers in the Empire State Building ran down dozens of flights of stairs and poured out into the streets. Air traffic control towers were evacu- ated at busy eastern U.S. airports, fouling the travel plans of thousands of people. Cell phone service was over- whelmed. The earthquake was felt south to Atlanta; north to Montreal and New Brunswick Province, Canada; and west to Detroit and Chicago. Damages totaled about $300 million and were suffered as far as Washington, D.C., where the Washington Monument and the National Cathedral both cracked, and in Brooklyn, New York. And yet, there were no deaths or serious injuries. Easterners got a taste of what westerners frequently experience.
Because seismic waves move slower than Internet traffic, some Twitter users in New York City and Boston read about the earthquake before they felt it. The citizen- based earthquake intensity website Did You Feel It? received more than 100,000 reports within four hours.
How Faults Work As our instrumentation and field equipment improve, we get better understanding of how faults work.
ELASTIC REBOUND The popular explanation of how faults move has been the elastic-rebound theory developed after the 1906 San Francisco earthquake. Based on surveyor’s measurements of ground along the San Andreas fault, it appears that Earth stresses cause deformation and movement on both sides of a fault ( figure 5.1 a, b ). However, the rocks along the fault itself do not move in response to this stress because they are rough and irregular, resulting in strong interlocking bonds with friction that retards movement. But as the land-masses away from the fault continue to move, energy builds up and is stored as elastic strain in the rocks. When the applied stresses become overpowering, the rocks at the fault rupture, and both sides quickly move forward to catch up, and even pass, the rocks away from the fault ( figure 5.1 c ). After a fault movement, all the elastic strain is removed from the area, and the buildup begins anew. The elastic-rebound theory is somewhat analogous to snapping a rubberband or twanging a guitar string; it even accounts for aftershocks. This idea has held sway for more than 100 years and is described in most textbooks. It still works as a first approx- imation to reality, but a better understanding has emerged in recent years.
NEWER VIEW Movements along a fault may be better visualized as win- dows of opportunity. Fault movement begins at a hypocenter
and then propagates outward for a certain distance and length of time. How much of the stored energy is released during an earthquake depends on the number of seconds the fault moves. For example, if there were 12 m (40 ft) of unreleased movement along a section of fault and the rupture event, passing by from front to end, lasted long enough for only 6 m (20 ft) of movement, then only half of the energy would have been released. An analogous event might be opening a locked gate to a long line of people. If the gate is held open only long enough for half the people to enter and is then closed and locked, the other people will simply have to wait until the next time the gate opens. This is an important modification of elastic-rebound theory. The elastic-rebound theory has said that after a big earthquake, most of the elastic strain is removed from the rocks, and considerable time will be required for it to build again to a high enough level to cre- ate another big earthquake. We no longer think this is true.
Another way to visualize how faults move is to imagine rolling out a large carpet to cover an auditorium floor. Sup- pose that the carpet misses covering the floor to the far wall by a foot. You can’t pull the rug the rest of the way to the wall; it won’t move because the friction is simply too great. How- ever, if you create a large ripple in the carpet and push the ripple across the auditorium floor, the carpet can be moved. Faults may act the same way. A small portion of a fault may slip, creating a ripple that concentrates elastic energy at its leading edge. The farther the ripple travels, the bigger the
Fault
(a)
(b)
(c)
Road
Figure 5.1 Elastic-rebound theory. (a) An active fault with a road as a reference line. (b) Deformation occurs along the fault, but friction of rock masses at the fault retards movement. (c) Finally, the deformation is so great that the fault ruptures, and the two sides race past each other and may actually catch up with and move past the earlier deformation.
abb22878_ch05_110-142.indd 111abb22878_ch05_110-142.indd 111 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
112 Chapter 5 Earthquakes Throughout the United States and Canada
earthquake. The moving ripple may encounter different amounts of unreleased energy in different areas of the fault.
Landers, California, 1992 New insight on how faults work was provided by the Landers area earthquakes in 1992 and 1999. This earthquake sequence began on 22 April 1992 with the right-lateral movement of the magnitude 6.1 Joshua Tree earthquake ( figure 5.2 ). Right-lateral movements along the fault trend resumed two months later at 4:58 a.m. on 28 June with the magnitude 7.3 Landers earthquake.
A third earthquake, triggered by the first two, broke loose a few hours later. At 8:04 a.m. on 28 June, the magni- tude 6.3 Big Bear earthquake came from a left-lateral move- ment that ruptured northeast toward the center of the Landers ground rupture. The ruptures of the 28 June earthquakes form a triangle, with the San Andreas fault as the base ( figure 5.2 ). These fault movements have acted to pull a triangle of crust away from the San Andreas fault, thus reducing the pressures that hold the fault together and keep it from slipping.
Activity continued along this trend on 16 October 1999 with the right-lateral movement of the Hector Mine earth- quake in a magnitude 7.1 event ( figure 5.2 ). Is this sequence of earthquakes finished? Probably not.
Examining the Landers earthquake records to see what happened during the 24 seconds that the faults moved 70 km (43 mi) can teach us a lot about how faults move:
1. Fault movements commonly are restricted to one fault, and the rupture front often stops at large bends or steps in
San Diego
San Bernardino
Palm Springs
Los Angeles
7.3MW Landers
1992
7.1MW Hector Mine
1999 6.3MW
Big Bear 1992
1987
Mexico
1987
1986
MojaveSan Andreas fault
Mtns. segment
6.1MW Joshua Tree
1992
Salton Sea
segment
segm ent
San Bernardino
Coachella Valley
Figure 5.2 Map of major earthquakes near the northern and southern ends of the Coachella Valley segment of the San Andreas fault. The triangular block of crust near the northern end has moved northward.
Landers
Epicenter
Right step pull apart
Right step pull apart
Movement stopped here
Em erson
Johnson V alley
Camp Rock
H om
estead Valley
0
0
5 mi
10 km
N
Figure 5.3 Northward-rupturing faults in the 1992 Landers earthquake. The rupture front slowed at right steps, and then moved onto adjacent faults before stopping in the middle of a straight segment.
the fault, thus ending the earthquake. The Landers earth- quake was different. It began right-lateral movement on the Johnson Valley fault and traveled northward about 20 km (12 mi) until reaching a right-step, pull-apart zone. The rupture front slowed, but it moved through the step and continued moving northward on successive faults for another 50 km (30 mi) until finally stopping within a straight segment of the Camp Rock fault ( figure 5.3 ).
2. Rupture velocity on the Johnson Valley fault was 3.6 km/sec (8,000 mph), slowing almost to a stop in the right step and then continuing northward at varying speeds.
3. The amount of slip on the faults varied from centimeters to 6.3 m (21 ft) along the fault lengths and below the ground. Figure 5.4 shows the movements calculated by seismologists Dave Wald and Tom Heaton. Look at their cross-section and visualize the fault in movement as the rupture front snaked its way northward, up, down, and not always involving all the fault surface.
4. Notice how the amount of fault movement at the ground surface differs from that at depth ( figure 5.4 ).
abb22878_ch05_110-142.indd 112abb22878_ch05_110-142.indd 112 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Thrust-Fault Earthquakes 113
5. Fault movement, and shaking, at the hypocenter was modest compared to what came later.
6. As the rupture front moved northward, only a small por- tion of the fault was slipping at any one time. Fault movement lasted 24 seconds, but the longest any one fault portion moved was less than 4 seconds.
7. Although the rupture front was slowed in a right step, the amount of slip behind the rupture front kept increas- ing until enough energy built up to cause movement through the step.
8. The earthquake triggered other earthquakes in Nevada, northern California, Utah, and Yellowstone Park, Wyoming. Fortunately for the Los Angeles megalopolis, the northward- moving fault directed its strongest seismic waves to the north into the sparsely inhabited desert. The triggered effects all occurred north of the northward-moving fault. This phenomenon is known as directivity, wherein a rupture moving along a fault sends more energy in the direction it is moving.
9. In the 1992 Landers earthquake, the faults moved from south to north; in the 1999 Hector Mine earthquake, it was the opposite, as the fault moved mostly from north to south.
10. Fault patches with little or no movement on figure 5.4 may become the origination points for future earthquakes.
Thrust-Fault Earthquakes Some damaging earthquakes result when compressional forces push one rock mass up and over another in a reverse- fault movement (see figure 3.10). Dip-slip faults of this type are also known as thrust faults, especially when the fault surface is inclined at a shallow angle. Many of these thrust faults do not reach the ground surface; they are called blind thrusts.
VIRGINIA, 2011: ANCIENT FAULTS CAN REACTIVATE The 23 August 2011 5.8M w event in Virginia occurred as a reverse-fault movement along a north-northeast striking fault about 10 km (6 mi) long. The hypocenter was only 6 km (3.7 mi) deep, which allowed seismic waves to reach the surface largely unweakened. The earthquake shook loose within the Central Virginia Seismic Zone, which has the same length and width, about 120 km (75 mi). Small earthquakes are frequent in this zone. Moderate-size events include a 4.5 M w event on 9 December 2003 and a ˜4.8M in 1875.
We like to explain earthquakes using plate-tectonic pro- cesses. But eastern North America does not have active tectonic- plate edges now, but it has a plate-tectonic past. More than 480 million years ago, continent collision, which included much reverse faulting, began building the Appalachian Mountains. This was part of the assembly of the superconti- nent Pangaea (see figure 2.25). More collisions followed, but by 220 million years ago Pangaea was being torn apart in a process that included much tensional faulting. Some of these ancient faults may be reactivating due to modern regional stresses.
NORTHRIDGE, CALIFORNIA, 1994: COMPRESSION AT THE BIG BEND Monday, 17 January 1994, was a holiday celebrating the birth of Martin Luther King, Jr. But at 4:31 a.m., the thoughts of most of the 12 million people in the Los Angeles area were taken over by a 6.7 M w earthquake. One of the many thrust faults that underlie the San Fernando Valley, the Pico blind thrust, ruptured at 19 km (11.8 mi) depth and moved 3.5 m (11.5 ft) northward as it pushed up the south-dipping fault surface ( figure 5.5 ). Northridge and other cities setting on the upward-moving fault slab (hangingwall) were subjected to some of the most intense ground shaking ever recorded.
0
2.5
D ep
th (
km )
5
7.5
10
12.5
15 65 60 55 50 45 40 35
Right step
Right step
30 Distance along strike (km)
Homestead Valley fault
25 20
2
3
4
6 5
15 10 5 0 –5 –10
Johnson Valley fault
Hypocenter
Camp Rock/Emerson faults
Ground surface
North South
Rupture direction
Figure 5.4 Slip on faults varied from centimeters to 6.3 m (21 ft) during movements of the 1992 Landers earthquake. The contour interval of slip areas is 1 m. Source: Wald and Heaton in Seismological Society of America Bulletin 84:668–691, 1994.
abb22878_ch05_110-142.indd 113abb22878_ch05_110-142.indd 113 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
114 Chapter 5 Earthquakes Throughout the United States and Canada
Ground acceleration was as high as 1.8 g (180% of gravity) horizontally and 1.2 g vertically. (At 1.0 g vertical accelera- tion, unattached objects on the ground are thrown up into the air.) This intense shaking caused the widespread failure of buildings ( figure 5.6 ), parking garages (see figure 3.32), and bridges that killed 57 people, injured 9,000 more, and caused $40 billion in damages. The damages included the disabling of the world’s busiest freeway system, creating months of
North
Pico thrust fault
1994 hypocenter
1971 hypocenter
San Fernando thrust fault
Figure 5.5 Block diagram of thrust-fault movements that created the 1994 Northridge and 1971 San Fernando earthquakes. In 1994, the Pico blind thrust fault moved 3.5 m (11.5 ft) up to the north from a 19 km (11.8 mi) deep hypocenter. The cities “riding piggyback” on the upward- moving thrust plate experienced intense ground shaking. In 1971, a block moved up to the south on the San Fernando thrust fault from a 15 km (9.3 mi) deep hypocenter.
Figure 5.7 Active faults in southern California are shown in yellow. The southern end of the San Andreas fault is on the east side of the Salton Sea (lower right). Follow the San Andreas fault up and to the left where it exits the photo in the upper left corner. The photo shows the San Andreas in its “Big Bend” that Southern California pushes against, creating mountains and thrust-fault earthquakes. Source: JPL/NASA.
Figure 5.6 Collapse of Bullocks department store in Northridge Mall. Some of the rigid brick walls failed during ground movement. Photo by Kerry Sieh, Caltech.
problems for drivers (see figures 3.29 and 3.36). In terms of deaths, this earthquake can be viewed as a near miss due to its early morning occurrence. Analysis of the failed buildings indicates that an estimated 3,000 people would have died if the seism had occurred during working hours.
The 1994 Northridge event was similar to the 1971 San Fernando earthquake, which had a magnitude of 6.6 and killed 67 people (see chapter 3). In 1971, the movement was up a north-dipping thrust fault that abuts the blind thrust that moved in 1994 ( figure 5.5 ). In 1971, the energy was directed toward the city of Los Angeles, but in 1994, the energy was directed away from the city.
Southern California pushes northward against the “Big Bend” of the San Andreas fault, creating thrust faults that are mostly east-west oriented ( figure 5.7 ). Satellite measure- ments of ground movement using the global positioning system (GPS) tell us that the Los Angeles region is experi- encing a compressive shortening of 10 to 15 mm/yr. The measured deformation could generate an earthquake with a magnitude in the mid-6s every six years, plus a seism of magnitude 7 every 10 years.
The 1971 and 1994 earthquakes may be omens for a more earthquake-active 21st century. The death and destruc- tion from numerous magnitude 6.5 to 7 earthquakes on thrust faults within the city of Los Angeles would exceed the prob- lems caused by a magnitude 8 event on the San Andreas fault some 50 to 100 km away.
SEATTLE, WASHINGTON The Seattle fault zone is oriented east-west and runs along the south side of Interstate 90 through the city of Seattle ( figure 5.8 ). The fault zone is 4 to 6 km (2.5 to 3.7 mi) wide
abb22878_ch05_110-142.indd 114abb22878_ch05_110-142.indd 114 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Normal-Fault Earthquakes 115
and has three or more south-dipping reverse faults. A major fault movement occurred there about 1,100 years ago, as indicated by the following evidence: (1) The former shore- line at Restoration Point was uplifted about 7 m (23 ft) above the high-tide line in a single fault movement. This earthquake appears to have had a magnitude of around 7, about the size of the 1989 World Series event in the San Francisco Bay area. (2) Numerous large landslides occurred at this time, includ- ing some that carried trees in upright growth position to the bottom of Lake Washington. The age of these trees was determined by carbon-14 dating. (3) Several tsunami depos- its have been recognized in the sediment layers of the area. Logs and trunks carried or buried by these large waves date to the same time period. (4) The same date appears in the ages of six major rock avalanches in the Olympic Moun- tains. The avalanches apparently were shaken into action by the earthquake. (5) Coarse sediment layers on the bottom of Lake Washington were formed by downslope movement and redeposition of sediment in deeper waters. These distinctive deposits appear to have been caused by the same earthquake.
Part of Seattle sits on a 10 km (6+ mi) deep basin filled with soft sediments that shake severely during an earthquake. Seattle residents were reminded of this earthquake hazard on 3 May 1996, when a magnitude 5.4 seism struck northeast of the city. Shaking in downtown Seattle’s Kingdome was intense enough in the seventh inning of a baseball game between the Seattle Mariners and Cleveland Indians to cause postponement of the game. The owner of the Seattle Mariners then tried to use the earthquake as justification for breaking his lease with the Kingdome. (The team now plays in a new stadium.) When the next major earthquake (greater than magnitude 6.5) occurs on the Seattle fault, it may cause
stunning levels of death and destruction. There are about 80 bridges and 1,000 unreinforced masonry (URM) build- ings that could suffer damages plus a tsunami 2 m (6.5 ft) high could be created.
Normal-Fault Earthquakes Some damaging earthquakes result when tensional forces pull one rock mass apart and down from another in normal- fault movements (see figure 3.9).
PUGET SOUND, WASHINGTON, 1949, 1965, 2001: SUBDUCTING PLATES CAN CRACK In recent decades, normal-fault movements have brought seismic jolts to cities in the Puget Lowlands ( figure 5.8 ). Three of these significant earthquakes were caused by down-to-the-east movements within the subducting Juan de Fuca plate.
At 11:55 a.m. on 13 April 1949, a jolt arose from a normal-fault movement 54 km (34 mi) below the Tacoma- Olympia area. The surface wave magnitude was 7.1M s , and eight people lost their lives. It could have been worse, since it happened during the day and badly damaged many schools, but luckily, it was the week of spring vacation, so the schools were largely vacant.
At 7:28 a.m. on 29 April 1965, the plate ruptured again— this time at 60 km (37 mi) depth below the Tacoma-Seattle area. The 6.5M s seism killed seven people. In 2010 dollars, the destruction totaled $315 million in 1949 and $105 million in 1965.
At 10:54 a.m. on Wednesday, 28 February 2001, a normal-fault earthquake radiated out from a hypocenter 52 km (32 mi) below the Tacoma-Olympia area. The magnitude 6.8 event shook more than 3 million residents of the Puget Sound for 45 seconds. In Olympia, the earthquake cracked the dome of the State Capitol and made the legisla- tors’ offices unusable and the governor’s home uninhabit- able. In Seattle, 30 people were caught on top of the swaying Space Needle, bricks fell from the Starbucks headquarters building onto parked cars, and Bill Gates’s talk at a hotel was interrupted as overhead lights crashed to the floor and fright- ened people who knocked down others in their hurry to get outside. The earthquake killed no one, injured about 400 people, and caused about $2 billion in damages.
In each case, settling of soft sediments and artificial fill during the shaking caused major problems for structures built on them. There was substantial damage to older masonry buildings with inferior mortar and to buildings with inade- quate ties between vertical and horizontal elements. Split- level homes suffered more than their share of damage as their different sections vibrated at different frequencies, helping tear them apart.
**
**
1996 epicenter
Seattle
fault
1965 epicenter
1949 epicenter2001
epicenter
Tacoma
Everett
Seattle
Olympia
Restoration Point
Puget Sound
Olympic Mountains
Up
Down Lake Washington
N
0 25 km
*
Figure 5.8 Map of the Puget Sound area. Up and down refer to movements on the Seattle fault.
abb22878_ch05_110-142.indd 115abb22878_ch05_110-142.indd 115 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
116 Chapter 5 Earthquakes Throughout the United States and Canada
For all the damage the 2001 earthquake caused, the dam- age it did not do is even more significant. Following the 1965 earthquake, Washington improved its building codes and made many structural changes, such as tying homes to their foundations more securely, removing water-storage tanks from the tops of school buildings, and strengthening more than 300 highway bridges. These investments were more than repaid in damages prevented and lives saved during the 2001 earthquake.
Deep Earthquakes Beneath the Puget Sound Primary emphasis on earthquake hazards in the Pacific Northwest has been focused on the subducting plates. The hypocenters in the 1949, 1965, and 2001 events were within the subducting plate at depth. Earthquakes 30 to 70 km (20 to 45 mi) deep occur beneath the Puget Sound about every 30 years. The subducting Juan de Fuca plate is only 10 to 15 million years old and is warm and buoyant. As the plate is pulled eastward, it reaches greater depths. The increasing temperature and pressure with depth cause the minerals mak- ing up the plate to become more dense and shrink. This builds up stresses that cause the plate to rupture, producing earthquakes with magnitudes as large as 7.5.
Neotectonics and Paleoseismology Geologic history plays out on a longer timescale than human history. An active fault may have a large earth- quake only once in several generations. How can the earthquake record be extended back further than the written historic record? Earthquake history can be read in sediments using the tech- niques of neotectonics ( neo means “young”) and paleoseismology ( paleo means “ancient”).
Faults slash through the land with compressive bends that cause land to uplift and pull-apart bends that cause land to drop down ( figures 5.9 and 5.10 ). The down-dropped or fault-dammed areas within the fault zone can become sites of ponds, receiving (1) sand wash- ing in from heavy rains, (2) clays slowly settling from suspension in ponded water, and (3) vegetation that lives, dies, and is buried by clay and sand. These processes produce a delicate record of sediment layers that may be disturbed and offset by later fault movements. This is a record we can read.
Older, more deeply buried layers have existed longer ( figure 5.11 ) and
Sag pond
Figure 5.9 A close-up of the San Andreas fault at Wallace Creek in the Carrizo Plain. Notice the offset streams and the ponded depressions formed at the fault. Have the movements been right or left lateral? Photo courtesy of Pat Abbott.
Linear valley
Linear valley or troughScarp Scarp
Bench
Spring
Offset drainage channel
Offset drainage channel
Sag pond
Linear ridge
Shutter ridge
Older fault traces
Newer fault traces
Figure 5.10 Schematic diagram of topography along the San Andreas fault in the Carrizo Plain. Notice the sag pond here and in figure 5.9 . Sediments deposited in these depressions allow the prehistoric record of earthquakes to be read. Source: Misc. Geol. Invest., US Geological Survey.
have been offset by more earthquake-generating fault move- ments. The amount of fault offset is proportional to an earth- quake’s magnitude; the greater the offset of sediment layers, the bigger the earthquake. These principles suggest a method to determine the approximate sizes of prehistoric earth- quakes. Simply dig a trench through the sediment infill of a fault-created pond and read the fault offsets recorded in the
abb22878_ch05_110-142.indd 116abb22878_ch05_110-142.indd 116 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Neotectonics and Paleoseismology 117
Sand
Sand
Bedrock
Bedrock
Sand
Clay
Clay
Clay
Coal (C14date = 240 years)
Coal (C14date = 500 years)
Coal (C14date = 745 years)
Figure 5.11 Schematic cross-section of trench wall across a fault-created pond. The fault offsets the once-continuous sediment layers. Notice that an upper layer of organic material formed 240 years ago is unbroken. At depth, a 500-year-old organic-rich layer has been offset. Deeper still, a 745-year-old organic-rich layer has been offset twice as much, indicating two major fault movements since it formed. What is the approximate recurrence interval between earthquakes at this site? When might the next big earthquake be expected here?
Figure 5.12 A trench wall across the San Andreas fault at Pallett Creek. Sandy layers are whitish, clay-rich layers are grayish, and organic-rich layers are black. The black layer in the center formed about 1500 CE . It has been offset 1.5 m (5 ft) horizontally and 30 cm (1 ft) vertically since 1500 CE . Photo courtesy of Pat Abbott.
Figure 5.13 Maze of trenches dug to determine the offset of a gravel-filled stream channel by the Rose Canyon fault. The offset here is 10 m (33 ft), and the fault is active. Photo by Pat Abbott.
sediments ( figures 5.11 and 5.12 ). Sediment layers in trench walls can be traced by digging a network of intersecting trenches to gain a three-dimensional view of fault offsets through time ( figure 5.13 ).
Dates of prehistoric earthquakes can be obtained by analyzing amounts of radioactive carbon in organic material (e.g., logs, twigs, leaves, and coal ) in the sediment layers. All life uses carbon as a fundamental building block. Most carbon occurs in the isotope C12, but a small percentage is
radioactive carbon (C14) produced in the atmosphere by bombardment of nitrogen atoms with subatomic particles emitted from the Sun. Carbon is held in abundance in the atmosphere as carbon dioxide (CO 2 ). All plants and animals draw in atmospheric CO 2 , and their wood, leaves, bones, shells, teeth, etc., are partly built with radioactive carbon. As long as an organism lives, it exchanges carbon dioxide with the atmosphere via photosynthesis or breathing. The per- centage of radioactive carbon in a plant or animal is the same as that of the atmosphere during the organism’s lifetime. However, when an organism dies, it ceases taking in
abb22878_ch05_110-142.indd 117abb22878_ch05_110-142.indd 117 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
118 Chapter 5 Earthquakes Throughout the United States and Canada
radioactive carbon, and the radiocarbon in its dead tissues decays with a half-life of 5,730 years. The presence of organic material allows us to determine the time of death and hence the age of enclosing sediments. This places actual dates into faulted sedimentary layers. The determination of real dates allows us to estimate the recurrence intervals for earthquakes—that is, how many years pass between earth- quakes at a given site.
The half-life of C 14 is short, thus restricting its usage to the last 50,000 years or so. This short half-life is useful for determining events in human history.
Figure 5.11 is a schematic representation of a trench- wall exposure of faulted pond sediments, demonstrating how fault-rupture sizes and recurrence intervals may be deter- mined. A real example of a faulted pile of ponded sediments is shown in figure 5.12. Here at Pallett Creek along the San Andreas fault, Caltech geologist Kerry Sieh has determined that fault movements with 6 m (20 ft) of horizontal offset recur about every 132 years. However, these 7+ magnitude earthquakes have occurred as close together as 44 years and as far apart as 330 years.
Earthquake Prediction The public really wants to have earthquakes foretold in much the same style and accuracy as they receive with weather forecasts. Our ability to forecast earthquakes on longer timescales is fairly good, but on short timescales we have no ability at all.
LONGTERM FORECASTS Can we predict earthquakes on intermediate to long time- scales using the paleoseismology approach? It seems to work well for some faults but not for others. Geologist Thomas K. Rockwell classifies fault-movement timing into three groups:
1. Quasi-periodic movements . These faults have major movements at roughly equal time intervals. This regular pattern can be defined using the trenching and radiocar- bon dating of paleoseismology.
2. Clustered movements . Adjacent fault segments move dur- ing several decades, and then they cease movement for a century or millennium until the next cluster begins. A good example of clustered movements is occurring right now on the North Anatolian fault in Turkey (see figure 4.26).
3. Random movements . These faults are inherently unpre- dictable; they have no definable pattern for their major movements. The San Andreas fault seems to be in this category.
In December 1988, using paleoseismologic analysis, a group of geologists forecast earthquake sizes and probabili- ties for some major faults in California ( figure 5.14 ). They placed a 30% probability on a magnitude 6.5 earthquake occurring on the Loma Prieta segment of the San Andreas fault within 30 years. Ten months later, the magnitude 6.9 World Series earthquake occurred there.
In 2003, the working groups stated that there is a 62% probability of at least one magnitude 6.7 earthquake striking
San Francisco Peninsula Parkfield
Fresno
North Coast
San Francisco
San Luis Obispo
Los Angeles
San Diego
Cholame Mojave
San Bernardino
Valley
Southern East Bay
Southern Santa Cruz Mountains
San Bernardino Mountains
Anza
San Jacinto Valley
Imperial Coachella
Valley
Borrego Mountain
Imperial fault
San Jacinto fault
San Andreas fault
Northern East Bay
Centralcreeping Carrizo
Hayward fault
<1 0%
<1 0%
32 %
20 %20
%
30 %
30 %
18 %
30 %
27 %
17 %
40 % 40
%
30 % 5
0%
90 %
77
77 77
77
77
77
77
6.56.588
88 7.57.5
7.57.5 7.57.5 6.56.5
<6.5<6.5
66
Figure 5.14 Working group analyses of expected earthquake magnitudes and their probabilities of occurring before the year 2032. Forecasts are based on historic records and trench-wall offsets of sediments dated by radiocarbon analyses and global positioning system measurements.
abb22878_ch05_110-142.indd 118abb22878_ch05_110-142.indd 118 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Earthquake Prediction 119
the San Francisco Bay region before 2032 and an 85% prob- ability of a magnitude 7 or higher earthquake in southern California before 2024.
SHORTTERM FORECASTS Our knowledge of earthquakes is quite impressive. Plate tectonics tells us why and where they occur, mostly along plate edges. Neotectonic analysis allows us to know how big and how often earthquakes have occurred on any fault. How- ever, many people are not satisfied; they want short-term prediction for earthquakes. Unfortunately, we are not even close to having that capability. We don’t have a workable theory, and it seems quite possible that the detailed behavior of faults is too unpredictable to ever allow short-term predic- tion of earthquakes. Theories of earthquake prediction that seem logical have been developed, and they still receive coverage in textbooks, but all of them have been proved false. Science is a demanding thought process. Beautiful ideas may have no substance. Creative hypotheses may have no validity. The truth is elusive.
A public eager for short-term prediction of earthquakes includes many gullible people. In 1977, Charles Richter commented that “journalists and the general public rush to any suggestion of earthquake prediction like hogs toward a full trough . . . [Prediction] provides a happy hunting ground for amateurs, cranks, and outright publicity-seeking fakers.”
Earthquake Weather In some regions, there are people who believe that earth- quakes are related to certain weather conditions, known as earthquake weather . The idea that earthquakes are related to weather is flawed. There is no connection between earth- quake energy released by fault movements miles below ground and the weather, which is due to solar energy received at Earth’s surface. Earthquakes are powered by the outflow of Earth’s internal energy; this is not affected by whether it is hot or cold, dry or humid, day or night, or any other weather condition.
Nostradamus Much ballyhoo surrounds the rhymed prophecies published in 1555 by the French doctor Michel de Notredame (Nos- tradamus). Vaguely worded statements by Nostradamus are believed by some people to predict earthquakes in our time. At the risk of being rude, the prophecies appear as truth only to undisciplined minds unable or unwilling to sort fact from fiction.
New Madrid, Missouri An early 1990s prediction event occurred when a dying economist named Iben Browning filled his final days with personal excitement by predicting a major earthquake in the mid-United States similar to the earthquakes of 1811–1812. Scientists could readily see that his predictions were based on an old failed hypothesis, but an uncritical print and elec-
tronic media went on a binge of emotional coverage as a horde of television crews and reporters descended on New Madrid, Missouri, eagerly awaiting the earthquake that never came.
Psychic Predictions Every so often we hear a psychic predict that a gigantic earthquake will cause California to break off and sink into the Pacific Ocean. Is this possible? No! This gigantic rupture-and-sink process is impossible; it is fantasy. Remember isostasy? Continents are made of less-dense rocks that float on top of denser mantle rocks. In fact, California did break off; it happened 5.5 million years ago as the Gulf of California began forming. California did not sink then and it won’t sink in the future. The faulted slice of western California and Baja California will continue moving northwest toward a rendezvous with Alaska. If present trends continue, in a few tens of million years, the Californias will plow into Alaska and become part of its southern margin. Southern California will switch from surfing beaches to ski slopes.
Experiment at Parkfield, California The U.S. Geological Survey forecasted a magnitude 6 earth- quake on the San Andreas fault in the Parkfield area based on the pattern of historical seismicity. Parkfield experienced magnitude 5.5 to 6 earthquakes six times in the historical period—in 1857, 1881, 1901, 1922, 1934, and 1966. Some people perceived the pattern of an earthquake about every 22 years. U.S. Geological Survey scientists forecast that the next earthquake would occur in 1988, plus or minus five years. Thus, in 1984, the Parkfield Prediction Experiment was launched by deploying an unprecedented array of instruments in the field with a large team of scientists to interpret every detail of the earthquake that would come by January 1993. Breaking news: It finally happened! A mag- nitude 6.0 earthquake occurred on 28 September 2004, 16 years after the forecast date. With more than 22 years of work and tens of millions of taxpayer dollars spent, the earthquake was unpredicted. The Parkfield Earthquake Experiment, the best-staffed and best-funded earthquake prediction experiment ever, was a total failure at short-term earthquake prediction.
What is our current understanding of the possibilities of short-term predictions of fault movements? First, there is no reason the fault-rupture process must occur with any regular- ity or predictability. Second, although it may not be hopeless to look for precursors to earthquakes, there clearly is more to earthquake triggering than can be explained simply by the steady loading of plate-tectonic stress onto faults that then rupture in evenly spaced, characteristic earthquakes. The bottom line for each person is this: short-term prediction of earthquakes is not forthcoming, so plan your life accord- ingly. Organize your home and office to withstand the big- gest earthquake possible in your area, and then don’t worry about when that day will come.
abb22878_ch05_110-142.indd 119abb22878_ch05_110-142.indd 119 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
120 Chapter 5 Earthquakes Throughout the United States and Canada
PERILS OF PREDICTION: SCIENTISTS ON TRIAL At 3:32 am Monday, 6 April 2009, the city of L’Aquila, Italy was hit by a 6.3 M earthquake that killed 309 people. L’Aquila is, or was, a charming medieval city with hundreds of old, fragile masonry buildings sitting in a seismically active area in Italy. Twice before, in 1461 and 1703, the city was largely destroyed by earthquakes. The difference this time is that four scientists and three engineers (one a government offi- cial) have been charged with felony manslaughter for their roles in reassuring the public that there was no significant earthquake danger. Their trial began before a single judge in late September 2011.
What events led to this unprecedented trial? Beginning in October 2008, dozens of low magnitude seisms hit the city and surrounding region. The first quarter of 2009 brought hundreds more small seisms. On 30 March, a 4.1 M seism shook up residents causing a government official to convene a meeting on 31 March of the Major Risks Committee in L’Aquila featuring the seven experts. After the meeting, the government official tried to calm residents fears of a big earth- quake including saying scientifically false statements such as “…it’s a favorable situation because of the continuous dis- charge of energy.” Townsfolk are reported to have comforted themselves by telling each other – the more tremors, the less danger. Some residents cancelled their evacuation plans and stayed in their homes – only to have family members die when their houses collapsed during the big earthquake. Residents felt betrayed. They pressed for legal action. The prosecution did not charge the commission members with failing to pre- dict the earthquake, but with presenting a superficial risk assessment with scientifically inaccurate information that gave false reassurances to the public. The prosecution asked for prison terms of four years; on 22 October 2012, the judge gave each defendant a six-year prison term.
Human-Triggered Earthquakes We humans trigger earthquakes in a variety of ways.
DISPOSAL WELLS Pumping liquids underground under high pressure can trig- ger earthquakes. This cause-and-effect relationship has long been known in the petroleum industry, but was dramatically proven in the 1960s in the Denver area.
Denver, Colorado In early 1962, in secret, the Rocky Mountain Arsenal began pumping chemical warfare waste under pressure down a well into old rocks 3.7 km (2.3 mi) deep. Earthquakes began one month later and rose to more than 40 per month, causing alarm in Denver. Pumping stopped in September 1963 and earth- quakes became minimal in number and magnitude. Pumping
resumed in September 1964 and so did earthquakes. In 1967, three of the earthquakes exceeded 5M and the hue and cry forced pumping to stop. And the earthquakes stopped. The relationship is clear. Fluids pumped underground under pres- sure can be forced into any ancient faults that are present, add- ing stress and reducing friction and thereby causing them to begin moving again. The greater the amount of fluid pumped, the greater the number and magnitude of earthquakes.
Ashtabula Township, Ohio On 25 January 2001, Ashtabula Township in Ohio was rocked by a magnitude 4.5 earthquake. The shaking damaged 50 houses and businesses as ceiling tiles fell, plaster cracked, and gas lines ruptured, forcing people to evacuate. This earthquake was the biggest in a series that began on 13 July 1987 with a magnitude 3.8 event. Why did earthquakes begin and keep recurring in this industrial port city on the shores of Lake Erie? In 1986, a 1.8 km (1.1 mi) deep well was drilled to inject hazardous wastes underground. For seven years, beginning in 1986, millions of gallons of waste-carrying liquids were forced down the well under pressure; earth- quakes began in 1987. The pressurized fuilds pumped under- ground encounter faults at depth, causing movements big enough to do damage at the surface.
INCREASES IN OIL AND NATURAL GAS PRODUCTION We humans have a huge thirst for fossil fuels to power our industries, transportation, and personal lives. This has led to increased use of hydraulic fracturing , commonly called fracking , wherein liquids are pumped down wells under high pressure in order to fracture and crack open rocks. The fractured rocks yield much greater volumes of natural gas and oil from deep underground. There is no debate: Hydrau- lic fracturing has significantly increased fossil-fuel energy production in the United States and other countries. How- ever, there are environmental concerns that have led to the banning of fracking in New Jersey and France.
Dallas–Fort Worth, Texas Hydraulic fracturing and new techniques of horizontal drill- ing are yielding enormous volumes of natural gas from rocks that previously were too “tight.” For example, more than 200 wells drilled into the Barnett Shale (a tight mudstone) in the Dallas–Fort Worth area now yield huge volumes of natural gas. But earthquakes up to 3.3M began in 2008 and contin- ued into 2009. The earthquake source was traced to one well drilled near an ancient fault. Abandonment of this well stopped the earthquakes. Natural gas production continues through the many other wells.
DAM EARTHQUAKES The downwarping of the land beneath the filling Lake Mead triggered many small earthquakes beginning in 1935 (see figure 2.7). This is a common occurrence; build a dam
abb22878_ch05_110-142.indd 120abb22878_ch05_110-142.indd 120 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
California Earthquake Scenario 121
and impound a reservoir of water, and then earthquakes fol- low. First, impounding a reservoir causes the earth to sink isostatically. Second, water seeping through the floor of the reservoir flows slowly underground throughout the region pushed by the large body of reservoir water above it. The underground water moves downward and outward as an advancing front of high pressure that may reach a fault and cause it to move. As an analogy, visualize what makes the water flow through the pipes in your house. In most cases, the water comes from a higher-elevation water tank or reser- voir that pushes the water down through the pipes.
China, 2008 Monday, 12 May 2008, began peacefully, like so many other days near the Dragon’s Gate Mountains in Sichuan, China. The 15 million people of the region were busy at work, their schools were full of children, and the giant pandas were at home in the Wolong Nature Reserve. But at 2:28 p.m., the earth ruptured along the base of the moun- tains and ripped northeastward along the Longmenshan fault for 250 km (155 mi) for about two minutes. When the shaking stopped, about 87,500 people were dead and 5 million were homeless. This massive earthquake was caused by the ongoing collision of India pushing into Asia. This rupture was a mountain-building thrust-fault event; similar movements over millions of years have built the Dragon’s Gate Mountains.
Time of day is always a factor in earthquake deaths, and the timing of this seism was terrible. Many of the build- ings were made of brittle concrete with little support steel, and at 2:28 p.m. on a Monday, the badly built schools and office buildings were full of people, resulting in a high death toll. The loss of so many children in the collapsed schools was especially tragic for families because of China’s one-child policy.
Plate tectonics was the cause of the earthquake, but what was the trigger? A debate is in progress. A 156 m (512 ft) tall dam was built in 2005 to create the Zipingpu Reservoir. In 2008, only 2.5 years later, the reservoir held 900 million tons of water. The dam lies 500 m (1,600 ft) from the Longmenshan fault. The weight of the reservoir water, plus the pore pressure of the water seeping underground, beneath the reservoir, would have caused the land to warp downward. Was this the added stress that triggered the fault to move on 12 May 2008 rather than 100 or so years later?
BOMB BLASTS Underground nuclear explosions in Nevada have triggered earthquakes. Some of the atomic-bomb blasts released energy equivalent to a magnitude 5 earthquake. The bomb explosions triggered significant increases in earthquakes in their region during the 32 hours after the blast.
Any time the level of stress or pressure is changed on rocks below the ground, earthquakes are possible. We humans can cause or trigger earthquakes.
Earthquake-Shaking Maps Computers are being used to create maps of earthquake shak- ing in near-real time.
DID YOU FEEL IT? Upon feeling an earthquake, a common response is to turn on the TV, radio, or computer to learn what just happened. Now you can help by sharing your shaking experience via your computer. Go through the USGS earthquake website to reach it, or simply google Did You Feel It? Click on Report Unknown Event, enter your ZIP code, and answer the ques- tions about what you felt. In a matter of minutes, a Commu- nity Internet Intensity Map will show the intensities of shaking felt in affected ZIP-code areas. You can be an impor- tant part of creating these Mercalli intensity maps and can also learn about the earthquake via your own participation.
SHAKEMAPS The intensities of seismic shaking are now recorded by instruments, and the data are fed into a computer that gener- ates a ShakeMap ( figure 5.15 ). The ShakeMap for the North- ridge earthquake shows the effects of directivity, with most of the intense shaking occurring north of the north-moving fault. The ShakeMap also shows other areas of more intense shaking in land underlain by soft rocks.
The rock and sediment foundations beneath buildings may amplify seismic waves ( figure 5.16 ). Seismic waves travel fast and with less amplitude in hard rocks. When seis- mic waves pass into soft rock or loose sediment, they slow down, but their amplitudes increase, and thus the shaking increases. Much of Los Angeles is built on soft rocks that amplify seismic shaking. In some areas, the soft rocks are 10 km (6 mi) thick, and seismic shaking may be amplified five times.
California Earthquake Scenario An analysis of probabilities for earthquakes greater than magnitude 6.7 in California shows the event is most likely to occur next on the southern San Andreas fault. A model of this earthquake and its effects was constructed through 13 special studies and 6 expert panels. The scenario earthquake is a magnitude 7.8 event that first ruptures at 7.6 km (4.7 mi) depth next to the Salton Sea and then continues rupturing northward past Palm Springs and through San Bernardino for 300 km (185 mi) ( figure 5.17 ). The modeled earthquake kills 1,800 people, injures another 50,000, and causes $213 billion in damages. The best way to reduce these numbers is to pre- pare in advance for an earthquake like this—and preparation begins with education and preparedness exercises.
abb22878_ch05_110-142.indd 121abb22878_ch05_110-142.indd 121 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
122 Chapter 5 Earthquakes Throughout the United States and Canada
Fillmore
Castaic
Palmdale
Pasadena Burbank
Westwood
Culver City Los Angeles
Whittier
Downey
Lakewood Torrance
Hawthorne
Inglewood
Newhall
San Fernando
Northridge
Malibu
119°00’W
34°00’N
Perceived shaking
Not felt
None None None Very light Light Moderate Heavy Veryheavy Moderate/ Heavy
Weak Light Moderate Strong Very strong Severe Violent Extreme
Potential damage
Instrumental intensity
Peak VEL. (cm/s)
Peak ACCEL. (%g) <.17 .17-1.4 1.4-3.9 3.9-9.2 9.2-18 18-34
16-318.1-163.4-8.11.1-3.40.1-1.1>0.1
I II-III IV V VI VII VIII IX X+
31-60 60-116 >116
34-65 65-124 >124
34°30’N
118°30’W 118°00’W
Simi Valley
Thousand Oaks
EPICENTER
Camarillo
Santa Paula
2010 300
km
Figure 5.15 ShakeMap for the magnitude 6.7 North- ridge earthquake in 1994. The fault moved to the north, and the greatest shaking was north of the epicenter. The variations in intensity of shaking are due to distance and variations in rock foundations.
Figure 5.16 Amplification of ground motion during an earthquake in Los Angeles. Amplification is minimal in hard rocks (purple), significant in softer rocks (red), and greatest where the softer rocks are the thickest (yellow). Source: Ned Field, US Geological Survey.
abb22878_ch05_110-142.indd 122abb22878_ch05_110-142.indd 122 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Earthquakes in the United States and Canada 123
An earthquake similar to the scenario event on the San Andreas fault broke loose in Alaska on 3 November 2002, mostly on the Denali fault; both are right-lateral faults. The Alaskan earthquake began in the west with faults mov- ing to the east and then southeast for 140 seconds, rupturing the ground for 340 km (210 mi) in a magnitude 7.9 seism with ground offsets up to 8.8 m (29 ft). Compare this earth- quake to the San Andreas fault event of 9 January 1857 that moved southward for 130 seconds, rupturing the ground for 360 km (220 mi) in a magnitude 7.9 seism with ground off- sets up to 9.5 m (31 ft). The 2002 Alaskan fault rupture had significant directivity; it was like a seismic shotgun aimed southeast, triggering earthquake swarms up to 3,660 km (2,270 mi) away in Washington, Wyoming, California, and Utah. In Lake Union in Seattle, the water sloshed back and forth and damaged houseboats. The good news about the 2002 Alaskan earthquake is that it happened in a remote area and had minimal effects on people. But this earthquake is like the fabled Big One of California, which will directly affect millions of people. For example, the San Andreas fault literally runs through the backyards of some of the 3 million people in the San Bernardino area.
ANNUALIZED EARTHQUAKE LOSSES Although big earthquakes do not happen in the United States every year, we forecast potential future costs as annualized earthquake losses. Data on population, buildings, and shak- ing potential are analyzed in a software program called HAZUS. For the United States, $4.4 billion in annual earth- quake losses are projected. Southern California accounts for
almost half of the losses; to Los Angeles County alone are attributed more than $1 billion in losses each year.
GREAT SHAKEOUT EVENTS A valuable way to prepare for earthquakes is to practice your response during a virtual earthquake. The first major virtual event was an effort by scientists and emergency managers to involve the public and schools in the “Great Southern California Shakeout” at 10 a.m. on 13 November 2008; it involved 5.4 million people. The concept has now spread to include great shakeout events in all of California, New Zealand, Nevada, Guam, British Columbia, Oregon, Idaho, and 10 states in the central United States. More countries have events planned.
One of the tips given to participants is that, upon feeling an earthquake, the immediate response that usually works the best is to:
Drop, Cover, and Hold on. The most common hazard is having objects fall on or be thrown at you. The best response is to “make like a turtle” and dive under a heavy table or desk to create a protective shell; then hang onto its legs to keep your shell until the shaking stops. The best way to remember this strategy is to practice it now so you can react instantly when everything starts shaking.
Earthquakes in the United States and Canada Awareness is growing that destructive and death-dealing earthquakes are a widespread problem, not just something that happens in California. Figure 5.18 is a map centered on the United States showing epicenters of significiant earth- quakes during a 92-year-long period. Compare the epicenter locations with the earthquake hazards map of the United States ( Figure 5.19 ). Figure 5.20 is a map of the eight largest earthquakes in Canadian history. All these figures show that earthquakes cluster in certain areas.
In Alaska and California, earthquakes occur in such large numbers and large sizes that they tend to obscure the earthquake history of the rest of the United States. If Alaska and California are ignored, the list of 10 largest U.S. earth- quakes shows that major seisms occur in numerous states— 10 major earthquakes, 10 different states ( table 5.1 ).
The history of earthquakes in Canada also shows variety ( table 5.2 ). The list is dominated by events along the tectoni- cally active west coast of British Columbia, yet the list of 11 largest earthquakes involves four provinces.
An expanded look at the earthquake history of the United States shows that all 50 states are hit by earthquakes, and many of the states have large earthquakes ( table 5.3 ). At least 18 states have been rocked by magnitude 6 or greater earth- quakes; some of the older seisms may have been as big, but scientific records are lacking.
Figure 5.17 The big earthquake most likely to occur next in California is a magnitude 7.8 rupture on the southern San Andreas fault. Here it is assumed that the rupture will begin at the Salton Sea and move northward through Los Angeles. Colors indicate severity of ground shaking. Source: Shakeout.org
abb22878_ch05_110-142.indd 123abb22878_ch05_110-142.indd 123 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
124 Chapter 5 Earthquakes Throughout the United States and Canada
Highest hazard
64+
48–64
32–48
16–32
8–16
4–8
0–4
% g
Lowest hazard
Figure 5.19 Earthquake hazards in the conterminous United States. Color show horizontal shaking, as a percentage of acceleration of gravity, that have a 2% probability of being exceeded in 50 years. Source: US Geological Survey, 2008.
Figure 5.18 Epicenters of earthquakes in the United States, southern Canada, and northern Mexico, 1899–1990.
abb22878_ch05_110-142.indd 124abb22878_ch05_110-142.indd 124 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Western North America: Plate Boundary–Zone Earthquakes 125
Magnitude
5.0–5.9 6.0–6.9
1929 1918 1700
1946
1970 1949
1985
1933
19291663
km
0 500 1000
Figure 5.20 The eight largest earthquakes in Canadian history, 1660–2011, are shown by red circles. See table 5.2 for more data. Source: EarthquakesCanada.
TABLE 5.1 Ten Largest Earthquakes in the United States (excluding Alaska and California)
Magnitude Date Location 9.0 26 Jan 1700 Washington, Oregon—
Cascadia subduction
7.9 2 Apr 1868 Hawaii—Ka’u district
7.5 7 Feb 1812 Missouri—New Madrid
7.3 16 Dec 1811 Missouri, Arkansas— New Madrid
7.3 17 Aug 1959 Montana—Hebgen Lake
7.3 31 Aug 1886 South Carolina—Charleston
7.2 16 Dec 1954 Nevada—Dixie Valley
7.0 23 Jan 1812 Illinois—New Madrid zone
7.0 28 Oct 1983 Idaho—Borah Peak
6.8 28 Feb 2001 Washington—Nisqually
TABLE 5.2 Eleven Largest Earthquakes in Canada
Magnitude Date Location 9.0 26 Jan 1700 British Columbia—
Cascadia subduction
8.1 22 Aug 1949 British Columbia—Queen Charlotte Island
7.7 27 Oct 2012 British Columbia—Queen Charlotte Island
7.4 24 Jun 1970 British Columbia—Queen Charlotte Island
7.3 20 Nov 1933 Northwest Territories— Baffin Bay
7.3 23 Jun 1946 British Columbia— Vancouver Island
7.2 18 Nov 1929 Newfoundland—Grand Banks
7.0 26 May 1929 British Columbia—Queen Charlotte Island
7.0 5 Feb 1663 Quebec—Charlevoix
6.9 23 Dec 1985 Northwest Territories— Nahanni
6.9 6 Dec 1918 British Columbia— Vancouver Island
Source: EarthquakesCanada (2006).
The historic record shows that earthquakes are wide- spread, but when earthquake frequency is examined, a differ- ent picture emerges. The location of all U.S. earthquakes of magnitude 3.5 and higher during a 30-year period shows a marked asymmetry ( table 5.4 ). Alaska has 57% and California 23% of these earthquakes. Add in Hawaii and Nevada, and those four states received 91% of these earth- quakes. Eight states had none (Connecticut, Delaware, Flor- ida, Iowa, Maryland, North Dakota, Vermont, and Wisconsin).
A primary goal of the remainder of this chapter is to understand the large earthquakes in the United States and Canada that do not occur on the edge of a tectonic plate. We will examine specific earthquakes and their causes in regional settings: western United States and Canada under the influ- ence of plate tectonics and buoyancy forces; the stable (tec- tonically “inactive”) central and eastern United States and Canada; and finally, the relationship between earthquakes and volcanism in Hawaii.
Western North America: Plate Boundary–Zone Earthquakes Much of the earthquake hazard in western North America is due to the ongoing subduction of small plates, as well as the continuing effects of the overridden, but not forgotten, Farallon plate. When considering the size of the Pacific, North American, and Farallon plates, it is easy to appreciate why earthquakes affect the entirety of western North America. Consider that the Pacific plate is more than 13,000 km (8,000 mi) across and that it is grinding past the North American plate, which is more than 10,000 km (6,250 mi) wide. How broad a zone is affected by these
abb22878_ch05_110-142.indd 125abb22878_ch05_110-142.indd 125 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
126 Chapter 5 Earthquakes Throughout the United States and Canada
passing giants? The affected zone must be large—as big as the entirety of western North America. The scale of these gigantic plates strongly suggests that their interactions are an underlying cause of earthquakes throughout the western United States, Canada, and Mexico.
WESTERN GREAT BASIN: EASTERN CALIFORNIA, WESTERN NEVADA Owens Valley, California, 1872 The famous naturalist John Muir was in his cabin in Yosemite Valley when:
At half past two o’clock of a moon-lit morning in March, I was awakened by a tremendous earthquake, and though I had never before enjoyed a storm of this sort, the strange thrilling motion could not be mistaken, and I ran out of my cabin, both
glad and frightened, shouting, “A noble earthquake!” feeling sure I was going to learn something. The shocks were so vio- lent and varied, and succeeded one another so closely, that I had to balance myself carefully in walking as if on the deck of a ship among waves, and it seemed impossible that the high cliffs of the Valley could escape being shattered. In particular, I feared that the sheer-fronted Sentinel Rock, towering above my cabin, would be shaken down, and I took shelter back of a large yellow pine, hoping that it might protect me from at least the smaller outbounding boulders. For a minute or two the shocks became more and more violent—flashing horizontal thrusts mixed with a few twists and battering, explosive, upheaving jolts—as if Nature were wrecking her Yosemite temple, and getting ready to build a still better one.
What happened on 26 March 1872? The fault zone on the western side of the Owens Valley broke loose along a length of 160 km (100 mi). This is the third longest fault
TABLE 5.3 Largest Earthquakes by State
State Date Magnitude or Intensity
Alabama 18 Oct 1916 5.1
Alaska 27 Mar 1964 9.2
Arizona 21 Jul 1959 5.6
Arkansas 16 Dec 1811 7.0
California 9 Jan 1857 7.9
Colorado 8 Nov 1882 6.6
Connecticut 16 May 1791 VII
Delaware 9 Oct 1871 VII
Florida 13 Jan 1879 VI
Georgia 5 Mar 1914 4.5
Hawaii 2 Apr 1868 7.9
Idaho 28 Oct 1983 7.0
Illinois 23 Jan 1812 7.0
Indiana 27 Sep 1909 5.1
Iowa 13 Apr 1905 V
Kansas 24 Apr 1867 5.1
Kentucky 27 Jul 1980 5.2
Louisiana 19 Oct 1930 4.2
Maine 21 Mar 1904 5.1
Maryland 16 Jul 2010 3.4
Massachusetts 18 Nov 1755 6.3
Michigan 10 Aug 1947 4.6
Minnesota 9 Jul 1975 4.6
Mississippi 17 Dec 1931 4.6
Missouri 7 Feb 1812 7.5
State Date Magnitude or Intensity
Montana 17 Aug 1959 7.3
Nebraska 28 Mar 1964 5.1
Nevada 16 Dec 1954 7.2
New Hampshire 24 Dec 1940 5.5
New Jersey 30 Nov 1783 5.3
New Mexico 15 Nov 1906 VII
New York 5 Sep 1944 6
North Carolina 21 Feb 1916 5.2
North Dakota 16 May 1909 5.5
Ohio 9 Mar 1937 5.4
Oklahoma 6 Nov 2011 5.6
Oregon 5 Aug 1910 6.8
Pennsylvania 25 Sep 1998 5.2
Rhode Island 11 Mar 1976 3.5
South Carolina 31 Aug 1886 7.3
South Dakota 2 Jun 1911 4.5
Tennessee 17 Aug 1865 5.0
Texas 16 Aug 1931 5.8
Utah 12 Mar 1934 6.6
Vermont 10 Apr 1962 4.2
Virginia 31 May 1897 5.9
Washington 26 Jan 1700 9
West Virginia 20 Nov 1969 4.5
Wisconsin 6 May 1947 V
Wyoming 17 Aug 1959 6.5
Source: US Geological Survey.
abb22878_ch05_110-142.indd 126abb22878_ch05_110-142.indd 126 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Western North America: Plate Boundary–Zone Earthquakes 127
rupture in California history after the 1906 San Francisco and 1857 Fort Tejon events ( figure 5.21 ). Today, Highway 395 runs in a north-south direction, right along the faults. The 1872 faulted zone is up to 15 km (10 mi) wide, with vertical drops (normal faulting) of as much as 7 m (23 ft) and horizontal offsets (right lateral) up to 5 m (16 ft). The epicen- ter was near the town of Lone Pine, where 27 people, about 10% of the residents, were crushed to death in the collapse of their adobe (dried mud blocks) and stone houses. The seism is estimated to have had a magnitude of about 7.4. So, big earthquakes do happen far away from the coastal zone and the San Andreas fault.
The Western Great Basin Seismic Trend This earthquake belt runs through eastern California and western Nevada and has a recognizable line of epicenters (see figure 5.18 ) and faults ( figure 5.22 ). In historic time, Nevada has averaged one earthquake with a magnitude in the 6s per decade and one with a magnitude in the 7s every 27 years. Why so many earthquakes? In the last 30 million years, the region between the eastern Sierra Nevada in California and the Wasatch Mountain front in central Utah has expanded in an east-west direction, opening up by several hundred kilometers (figures 5.23 and 5.24). This extended area is known as the Great Basin, or the Basin and Range province. Nevada, in the heart of the extended province, has about doubled in width. As much as 20% of the relative motion between the Pacific and North American plates may be accommodated in the Basin and Range province. Extensional, pull-apart tectonics stretch the area, leaving numerous north-south- oriented, back-tilted mountain ranges separated by down-dropped, sediment-filled basins ( figure 5.23 ). The extension is accomplished with
TABLE 5.4 Most Active Earthquake States (magnitudes 3.5 and above, 1974–2003)
State Number of Earthquakes 1. Alaska 12,053
2. California 4,895
3. Hawaii 1,533
4. Nevada 778
5. Washington 424
6. Idaho 404
7. Wyoming 217
8. Montana 186
9. Utah 139
10. Oregon 73
Top 10 states total 20,702
Bottom 40 states total 378
Source: US Geological Survey. Highway
395
Fault
Fault
Figure 5.21 View to the north in Owens Valley. Faults are subparallel and to the left of Highway 395; note that the town of Lone Pine (in center of photo) is down-dropped. There is a lake in the right-stepping pull-apart between two fault segments. The Alabama Hills are at left center, and the Sierra Nevada in upper left. Photo by John S. Shelton.
normal faulting, so vertical separation dominates over hori- zontal slippage.
Some major earthquakes of historic times have occurred in the western part of the Great Basin province ( figure 5.22 ). (1) On 2 October 1915, a large earthquake occurred south of Winnemucca in Pleasant Valley, Nevada. This magnitude 7.7 event ruptured the surface for 59 km (37 mi). The slip was dominantly vertical (normal) with displacements up to 5.8 m (19 ft) ( figure 5.25 ). Some fault strands had right- lateral components of offset up to 2 m (6.5 ft). (2) On 21 December 1932, a magnitude 7.2 event occurred near Cedar Mountain, Nevada, rupturing the ground for 61 km (38 mi). (3) The year 1954 was a big one for earthquakes in Nevada. Events included a magnitude 6.6 on 6 July and a 6.9 on 24 August near Fallon, as well as two shocks of 7.2 and 6.9 that rocked Dixie Valley on 16 December. Figure 5.22 shows several gaps in the trend of historic, long ruptures of faults. Residents in these seismic gaps may be in for some surprises.
abb22878_ch05_110-142.indd 127abb22878_ch05_110-142.indd 127 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
128 Chapter 5 Earthquakes Throughout the United States and Canada
Southern Sierra gap
1952
zone
S ierra
1934 1932
1954 1903
White Mountain
gap
Stillwater gap
1915
Central Nevada
zone Trinity Range block
1869?
2008 Reno
1950
O re
g o
n Id
a h
o
Ruby Range
1934
1983
Intermountain seismic
belt
Utah ArizonaNevada
California
1872
E astern
37°
42°
120° 114° 1959
Key
Historic faulting No recent fault movements
Holocene faulting areas (past 11,000 ± years)
0
0 100 km
100 mi
Boundary of Great B
asi n p
rovi nce
Figure 5.22 Generalized map of historic faulting in the western Great Basin. Areas of ground broken by large earthquakes are in dark orange; notice the seismic gaps in the trend. Areas with numerous smaller seisms are brown.
Basin and
Range
Coast Ranges
Rocky Mountains
EW
Pacific Ocean
Great Valley
Sierra Nevada
Wasatch Mtns
Colorado Plateau
Park Range
Front Range
Figure 5.23 Schematic cross-section oriented west-east across the western United States. The Basin and Range province has stretched to double its initial width. This extension has created normal faults that generate earthquakes.
Reno, Nevada, 2008 Big earthquakes usually occur as a mainshock followed by numerous aftershocks. But sometimes quakes occur in a swarm, a cluster of earthquakes without a mainshock. Between
28 February and 3 June 2008, Reno experienced a swarm of 1,090 quakes of magnitude 2 and greater ( figure 5.22 ). The peak of the swarm occurred in late April and early May when the numbers of earthquakes increased and magnitudes reached
abb22878_ch05_110-142.indd 128abb22878_ch05_110-142.indd 128 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Western North America: Plate Boundary–Zone Earthquakes 129
Figure 5.25 A portion of the fault scarp created during the 1915 earthquake in Pleasant Valley, Nevada. Notice the people for scale (white arrows). Photo courtesy of Pat Abbott.
Figure 5.24 Computer-generated image of topography in the western United States. Notice in the center the north- south-oriented mountains separated by linear valleys. Basin and Range topography is outlined by the red line. Photo from US Geological Survey.
4.2, 4.7, 4.2, and 3.8. The swarm occurred along a short fault and may well have been an interval of fault growth in response to Nevada being pulled apart by tectonic forces. The good news for Reno residents is that a short, poorly developed fault will not produce a large earthquake. The bad news is that the Reno area has other longer faults with a 65% chance of pro- ducing a magnitude 6 earthquake in the next 50 years.
THE INTERMOUNTAIN SEISMIC BELT: UTAH, IDAHO, WYOMING, MONTANA The Intermountain seismic belt is a northerly trending zone at least 1,500 km (930 mi) long and about 100 to 200 km (60 to 125 mi) wide ( figure 5.26 ). The belt extends in a
Great Falls
Missoula
48�
46�
44�
42�
40�
38�
36� 116� 114� 112� 110� 108�
Butte
Montana Wyoming
Hebgen Lake
Idaho FallsBorah Peak
Pocatello Twin Falls
Idaho
Nevada Rock SpringsEvanston
Elko Salt Lake City Vernal
Moab
Lake Powell
Lake Mead
Great Salt Lake
Se vie
r D es
ert
Paradox Valley
Utah Colorado
Ely Delta
Arizona New Mexico
200 km1000
Flaming Gorge Reservoir
MAGNITUDES
2.5+
4.0+ 5.0+
6.0+
7.0+
Figure 5.26 Earthquake epicenters in the Intermountain seismic belt, 1900–1985.
abb22878_ch05_110-142.indd 129abb22878_ch05_110-142.indd 129 27/12/12 3:49 PM27/12/12 3:49 PM
G A R R E T T , M E G A N 1 3 2 4 T S
130 Chapter 5 Earthquakes Throughout the United States and Canada
curved pattern from southern Nevada and northern Arizona into northwestern Montana (see figure 5.22 ). In effect, the seismic belt is the eastern boundary of the extending Basin and Range province. The bounding faults on the eastern side of the Great Basin are mostly down-to-the-west, whereas the bounding faults on the western side (in eastern California and western Nevada) are mostly down-to-the-east. The earth- quakes reaffirm that this part of the world is being stretched and pulled apart.
Hebgen Lake, Montana, 1959 The Rocky Mountains in the summertime are a beautiful place to be. On the moonlit evening of 17 August 1959, campers were settled into their spots at the Rock Creek Campground at the foot of the high walls of the Madison River Canyon. But at 11:37 p.m., the ground shook, and then an odd wind blew briefly down the canyon at high velocity. The wind was cre- ated by the push of an enormous rock slide. The south wall of the canyon dropped 43 million cubic yards of rock, which slid down the steep slope, across the Madison River, and moved about 150 m (500 ft) up the north wall ( figure 5.27 ). It entombed 26 campers. The gigantic landslide buried the can- yon to depths of 67 m (220 ft) and created a natural dam that began trapping a large body of water—Earthquake Lake.
What caused this life-ending landslide? At Hebgen Lake, directly west of Yellowstone National Park, two subparallel faults (Hebgen and Red Canyon) moved within 5 seconds of each other with 6.3m b and 7.5M s events ( figure 5.26 ). These two normal faults had their southwestern sides drop 7 and 7.8 m (23 and 26 ft) down fault surfaces inclined 45° to 50° to the southwest. The fault movements created a huge seiche in Hebgen Lake.
Borah Peak, Idaho, 1983 Just after 7 a.m. on 28 October 1983, the Lost River fault broke free 16 km (10 mi) below the surface and ruptured
northwestward 0.45 m (1.5 ft) horizontally and 2.7 m (9 ft) vertically for a 7.3M s event (figures 5.26 and 5.28). When the fault finished moving, Borah Peak, Idaho’s highest point, was 0.3 m (1 ft) higher, and the floor of Thousand Springs Valley was several feet lower. The ground shaking caused Thousand Springs Valley to live up to its name as under- ground water, squeezed out by the subterranean pressures, spouted fountains 3 to 6 m (10 to 20 ft) high.
The Wasatch Fault In historic times, large seisms have occurred in eastern California and western Nevada on the west and in Montana and Idaho on the east, but not on long sections of faults in Utah. Over 80% of Utah’s population lives within sight of the scarps , or steep slopes, of the 370 km (230 mi) long Wasatch Front, the zone of normal faults separating the mountains from the Great Basin ( figure 5.29 ). No large earthquakes have been reported along the Wasatch Front faults since the arrival of Brigham Young in 1847, but the sharply defined
Figure 5.27 Madison Canyon landslide and resulting lake, caused by the earthquake of 17 August 1959. Photo taken 24 days later by John S. Shelton.
Figure 5.28 The 28 October 1983 Borah Peak earthquake was a 7.3M s event with 2.7 m (9 ft) of vertical offset. Notice that some left-lateral offset also occurred. Mount Borah (in background) was uplifted slightly by this event. Photo from NOAA.
abb22878_ch05_110-142.indd 130abb22878_ch05_110-142.indd 130 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Western North America: Plate Boundary–Zone Earthquakes 131
earthquake threat and the danger for their towns. The fault segments shown in figure 5.29 are each capable of events like those of Hebgen Lake and Borah Peak. In the last 6,000 years, a magnitude 6.5 or stronger earthquake has occurred about once every 350 years on one of the Wasatch system faults. Parts of Salt Lake City, Provo, and Ogden lie on soft lake sediments that will shake violently during a large seism. The fault segment near Brigham City has not moved in the last 2,400 years and is a likely candidate for a major event.
RIO GRANDE RIFT: NEW MEXICO, COLORADO, WESTERNMOST TEXAS, MEXICO The Rio Grande rift is one of the major continental rifts in the world. It is a series of interconnected, asymmetrical, fault- block valleys that extend for more than 1,000 km (620 mi) ( figure 5.31 ). Here, it appears that the continental crust is being heated from below and is stretching. The crust responds by thinning and extending with accompanying normal fault- ing. In the last 26 million years, about 8 km (5 mi) of crustal extension has occurred near Albuquerque, New Mexico, a rate of about 0.3 mm/yr. The dominant motion on the faults is vertical, and the offset totals 9 km (5.5 mi). The rift basin is strikingly deep in places, yet most of the vertical relief created by fault offsets has been lessened by the copious quantities of volcanic materials and sediments that have poured into the rift over millions of years.
The topographic trough of the rift valley has attracted a major river (Rio Grande), which in turn has enticed human
G re
at B
as in
W as
at ch
R an
ge
1600 CE
1400 CE
700 CE
1000 CE
400 BCE
Ogden
Great Salt Lake
Salt Lake City
Brigham City
Provo
Nephi
Idaho
Utah
W yo
m in
g
0
0 80 km
50 mi
Figure 5.29 Map of faults along the Wasatch Front, Utah. The Wasatch fault has several segments. Dates of the most recent magnitude 6.5 or greater seisms are shown.
Figure 5.30 Aerial view eastward over Salt Lake City to the Wasatch fault running along the base of the mountains. The Wasatch fault zone is colored red. Photo by John S. Shelton.
faults show obvious potential for earthquakes ( figure 5.30 ). In an 1883 article in the Salt Lake Tribune , the famous geologist G. K. Gilbert warned the people of Utah of the
Basin and Range province
Rio Grande rift
Figure 5.31 An east–west-oriented extension has pulled apart some of western North America to form the topography of Basin and Range province. The Rio Grande rift is a geologically youthful rift valley. The sediment-filled basins are shown in orange.
abb22878_ch05_110-142.indd 131abb22878_ch05_110-142.indd 131 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
132 Chapter 5 Earthquakes Throughout the United States and Canada
bursting of the waves, large fissures were formed, some of which closed again immediately, while others were of vari- ous widths, as much as thirty feet, and of various lengths. These fissures were generally parallel to each other, nearly north and south, but not all. In some cases instead of fissures extending for a considerable distance there were circular chasms, from five to thirty feet in diameter, around which were left sand and bituminous shale, which later would burn with a disagreeable sulphorous smell.
The region is composed of thick deposits of water- saturated, unconsolidated sands and muds dropped by the Mississippi River. These loose materials intensified the shak- ing of the earthquakes, and the weak sediments flowed like water, erupted as sand volcanoes, and in places quivered like Jell-O. Several long-lasting effects of the New Madrid earth- quakes can still be seen in the topography. A 240 km (150 mi) long area alongside the Mississippi River sank into a broadly depressed area, forming two new lakes: Lake St. Francis, 60 km (37 mi) long and 1 km (0.62 mi) wide; and Reelfoot Lake in Tennessee, 30 km (19 mi) long, 11 km (7 mi) wide, and up to 7 m (23 ft) deep. Reelfoot Lake, now a bird sanctu- ary, hosts the gray trunks of cypress trees drowned more than 200 years ago; they still stand as silent testimony to the area’s earth-wrenching events ( figure 5.33 ). Other topographic fea- tures created by the seisms include (1) long, low cliffs across the countryside and streams with new waterfalls up to 2 m (7 ft) high; (2) domes as high as 6 m (20 ft) and as long as 24 km (15 mi); and (3) former swamplands uplifted and transformed into aerated soils.
settlers in need of water. Today’s settlements include Albu- querque, Socorro, and Las Cruces in New Mexico, El Paso in Texas, and Ciudad Juarez in Mexico. Historic earthquakes have had only small to moderate magnitudes, but the conti- nental lithosphere continues to extend, thus presenting a real hazard for large earthquakes.
Intraplate Earthquakes: “Stable” Central United States The map of earthquake epicenters in the United States (see figure 5.18 ) shows that the western third of the country has an elevated level of seismic activity. But there are clusters of epicenters in the “stable” central and eastern United States, the intraplate regions away from the active plate edges. There are not as many epicenters, but some individual earthquakes are just as big. Seismic hazards are significant (see figure 5.19 ).
NEW MADRID, MISSOURI, 18111812 A succession of earthquakes rocked the sparsely settled cen- tral part of the Mississippi River Valley at the time of the War of 1812. Between 16 December 1811 and 15 March 1812, Jared Brooks, an amateur seismologist in Louisville, Kentucky, recorded 1,874 earthquakes. He classified eight of them as violent and another 10 as very severe. The four larg- est events occurred on 16 December 1811 (two), 23 January 1812, and 7February 1812. The hypocenters were located below the thick pile of sediments where the Mississippi and Ohio rivers come together, at the upper end of the great Mississippi River embayment ( figure 5.32 ). These major seisms are called the New Madrid earthquakes, taking their name from a Missouri town of 1,000 people. Although few people were killed, the destruction of ground and buildings at New Madrid tolled the end of its importance as “the Gateway to the West.”
The following is excerpted from an eyewitness account of a New Madrid earthquake:
Accompanying the noise, the whole land was moved and waved like waves of the sea, violently enough to throw per- sons off their feet, the waves attaining a height of several feet, and at the highest point would burst, throwing up large volumes of sand, water, and in some cases a black bitumi- nous shale, these being thrown to a considerable height, the extreme statements being forty feet, and to the top of the trees. With the explosions and bursting of the ground there were flashes, such as result from the explosion of gas, or from the passage of the electric fluid from one cloud to another, but no burning flames; there were also sulphuretted gases, which made the water unfit for use, and darkened the heav- ens, giving some the impression of its being steam, and so dense that no sunbeam could find its way through. With the
Mississippi embayment
Gulf of Mexico
A tla
nt ic
O ce
an
Coastal plain sediments
Figure 5.32 Map of coastal-plain sediments deposited by rivers eroding North America. The Mississippi River embayment contains a mass of soft sediments.
abb22878_ch05_110-142.indd 132abb22878_ch05_110-142.indd 132 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Intraplate Earthquakes: “Stable” Central United States 133
earthquakes was on 16 December 1811 and seems to have occurred on the Cottonwood Grove fault as 13 ft (4 m) of slip along a 37 mi (60 km) rupture length. This seism is likely to have triggered two ruptures on the Reelfoot blind-thrust fault at New Madrid, Missouri: one also on 16 December 1811 and the largest of the series on 7 February 1812. The large earthquake that occurred on 23 January 1812 has been the most difficult to locate. At present, it seems the earthquake
Felt Area The New Madrid earthquakes have never been equaled in the history of the United States for the number of closely spaced, large seisms and for the size of the felt area ( figure 5.34 ). The earthquakes were felt from Canada to the Gulf of Mexico and from the Rocky Mountains to the Atlantic seaboard, where clocks stopped, bells rang, and plaster cracked. These big earthquakes were not a freak occurrence. The oral history of the local American Indians tells of earlier dramatic events.
Assessments of the earthquakes based on felt area yield magnitude estimates of 8 to 8.3. But are the sizes of the felt areas in figure 5.34 a good indicator of earthquake magni- tude? Were the New Madrid seisms many times bigger than the 1906 San Francisco earthquake? Not necessarily. The size of the felt area is related to the types of rocks being vibrated. The New Madrid seisms shook the rigid basement rocks (more than 1 billion years old) of the continental inte- rior. They rang like a bell, and the seismic energy was trans- mitted efficiently and far. The San Francisco earthquake took place in younger, tectonically fractured rocks that quickly damped out the seismic energy, thus confining the shaking to a smaller area.
Magnitudes Can aftershocks from the 1811–1812 earthquakes still be occurring more than 200 years later? Yes, they happen in the low-stress, midcontinent region. Modern studies of the dis- tribution of small aftershocks are helping define the faults that ruptured in 1811–1812 ( figure 5.35 ). The first of four big
Figure 5.33 Trunks of cypress trees drowned in Reelfoot Lake after water was dammed by the New Madrid earthquakes. (These drowned trees are analogous to those drowned by the January 1700 Pacific Northwest magnitude 9 earthquake; see figure 4.17.) Photo from US Geological Survey.
1811 New Madrid
1906 San Francisco
1971 San Fernando 1886
Charleston
Gulf of Mexico
P acific O
cean
A tla
nt ic
O ce
an
Figure 5.34 Felt areas of some large earthquakes in the United States. Orange areas are Mercalli intensities greater than VII; red areas are intensities of VI to VII.
abb22878_ch05_110-142.indd 133abb22878_ch05_110-142.indd 133 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
134 Chapter 5 Earthquakes Throughout the United States and Canada
vibrations (remember the 1985 Mexico City and 1989 World Series events); and (4) a very large area will be subjected to strong shaking. The effects of a magnitude 7.5 earthquake could include deaths in the thousands and damages in the tens of billions of dollars.
How frequent are large earthquakes here? Paleoseismo- logic analyses of sediment and wood indicate major earth- quake clusters occurred around 2350 bce , 900 ce , 1450 ce , and 1812 ce . A magnitude 7 or greater earthquake could occur here about every 500 years. There is a 7–10% chance of one in the next 50 years. If there is good news, it is the low frequency of occurrence for these large earthquakes. The les- sons of history here must be learned, and all new construction in the region should be built to with stand major earthquakes.
The earthquake threat in this region also includes lesser events. In 1843, it was a magnitude 6.4 event at Marked Tree, Arkansas, and in 1895, it was a magnitude 6.8 at Charleston, Missouri ( figure 5.35 ). Paleoseismologic analyses in trenches cut across faults and folds has led the US Geological Survey to forecast a 25–40% chance of a magnitude 6 to 7 earth- quake here within the next 50 years. Remember, the 1994 Northridge earthquake was a magnitude 6.7, and it killed 57 people and caused $40 billion in damages. Although earth- quakes in the central United States have low probability, they have high impact.
Since New Madrid sits in the continental interior, away from the active plate edges, why do big earthquakes occur here? The answer is unresolved, but a look at the geologic history of the region provides some understanding.
REELFOOT RIFT: MISSOURI, ARKANSAS, TENNESSEE, KENTUCKY, ILLINOIS Figure 5.35 shows that the epicenters of the large earthquakes line up along the Mississippi River Valley. Figure 5.32 is a map of the southern and eastern United States depicting the distribution of coastal-plain sediments—the sands and muds dropped by rivers eroding the North American landmass. The Mississippi embayment stands out as a prominent feature. Why are these sediments deposited so much farther into con- tinental North America? Why does the sediment distribution parallel the epicenters? Is it a coincidence that this same linear pattern keeps reappearing? No. Is it random chance that the Mississippi River flows along the course that it does? No.
The results of studies of seismic waves, gravity, and magnetism define a linear structural feature in the basement rocks underlying the New Madrid region ( figure 5.37 ). There is a northeast-trending depression at depth that is more than 300 km (190 mi) long and about 70 km (43 mi) wide. It is linear, has nearly parallel sides, and is about 2 km (1.2 mi) deeper than the surrounding basement rocks. In short, it is an ancient rift valley, known as the Reelfoot rift, formed about 550 million years ago. Similar features that are still forming today and are more apparent at the Earth’s surface include the Rio Grande rift in New Mexico (see figure 5.31 ) and the
happened about 125 mi (200 km) to the north, around south- ern Illinois. If this interpretation holds up, the hazards associ- ated with midcontinent earthquakes are more widely distributed than is commonly recognized.
The defined fault-rupture lengths are too short to have generated magnitude 8 earthquakes as suggested from felt- area analysis. When moment magnitudes calculated from fault surface-area estimates are considered with Mercalli intensities, the earthquake magnitudes seem to range from about 7.3 to 7.7. However, remember that the earthquake epicenters sit on top of a thick pile of water-saturated sedi- ment. This loose material amplifies the local shaking several times, leading to high Mercalli intensities. Discounting for amplification of seismic waves leads to earthquake magni- tude estimates, in chronologic order, of 7.3, 7.0, 7.0, and 7.5.
The Future When large earthquakes return again to the upper Mississippi River region, the potential for death and destruction is stag- gering ( figure 5.36 ). (1) The area has a large population (e.g., Memphis, St. Louis, Nashville); (2) many buildings were not designed to withstand large seisms; (3) the wide extent and great thickness of soft sediments will amplify seismic
Illinois
Charleston
Crowleys Ridge
Arkansas
Missouri
Mississippi River
Tennessee Marked Tree
Reelfoot Lake
Memphis
K en
tu ck
y
New Madrid
7 Feb. 1812
23 Jan 1812
Co tto
nw oo
d G ro
ve16 Dec 1811
Cairo
38o
37o
91o 90o
N
Figure 5.35 Map of the New Madrid region showing epicenters of large earthquakes and faults located using aftershocks. Marked Tree, Arkansas, was the site of a magnitude 6.4 event in 1843; Charleston, Missouri, had a magnitude 6.8 seism in 1895. Faults are indicated by dashed lines.
abb22878_ch05_110-142.indd 134abb22878_ch05_110-142.indd 134 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Intraplate Earthquakes: “Stable” Central United States 135
Some faults that were close to failure have failed; other faults will follow.
ANCIENT RIFTS IN THE CENTRAL UNITED STATES It is the fate of all continents to be ripped apart from below. Continents are rifted and then drifted and reassembled in different patterns. Sometimes the rifting process stops before separating a continent. The Reelfoot rift, now occupied by the Mississippi River, is a prominent failed rift. Other failed rifts, from different plate- tectonic histories, also exist beneath the surface in North America ( figure 5.38 ).
Failed rifts remain as zones of weak- ness that may be reactivated by later plate- tectonic stresses to once again generate earthquakes. Because failed rifts are deeply buried, they are difficult to study. Yet they raise significant questions. What are the fre- quencies of their major earthquakes? In general, the recurrence intervals for major earthquakes appear to be from a few hun- dred to more than a thousand years. How great an earthquake might be produced at each rift? The New Madrid earthquake series offers a sobering benchmark. Approximately 83% of the large earth- quakes recorded in the central United States are at or near the sites of ancient rifts.
The buried, ancient rifts in Figure 5.38 correlate with active fault zones at the sur- face. There are several examples. (1) The
St. Louis arm corresponds to the Ste. Genevieve fault zone. (2) The Rough Creek rift is expressed as the Rough Creek fault zone, appearing to continue eastward as the Kentucky River fault zone. Trenches dug across the Rough Creek fault have exposed the sedimentary records of reverse-fault move- ments with 1.1 m (more than 3.5 ft) of offset. (3) The south- ern Oklahoma rift corresponds with the frontal-fault system of the Wichita Mountains. Although this zone does not gener- ate earthquakes at present, the land surface testifies to major earthquakes. The Meers fault is dramatic enough to make any Californian proud, but this fault strikes N 63°W across south- western Oklahoma. Its fault scarp is 5 m (16 ft) high and 27 km (17 mi) long, and it has left-lateral offset up to 25 m (82 ft). At least two major fault ruptures have occurred there in geologically recent time. (4) The southern Indiana arm is overlain by the Wabash Valley fault zone, which appears to connect with the New Madrid zone. Prehistoric earthquakes read in the sedimentary record suggest seisms with m b equal to 6.3 to more than 7. Damaging earthquakes occur in the area about once a decade. Examples include a magnitude 5.5 in November 1968, a magnitude 5.2 in June 1987, and a magnitude 5 in southwest Indiana on 18 June 2002.
East African Rift Valley (see figure 4.6). The ancient Reel- foot rift was filled and covered by younger sediments ( figure 5.37 ). Today, the opening Atlantic Ocean basin pushes North America to the west-southwest, and some of the ancient faults of the Reelfoot rift are being reactivated to produce the region’s earthquakes.
Isostatic Rebound as an Earthquake Trigger Although stress from plate-tectonic movements elastically strains the rocks, what triggers the earthquakes? A new hypothesis by Purdue University geologist Eric Calais and colleagues suggests that isostatic rebound may be the trigger for movement of the ancient faults. As the North American ice sheet retreated northward from the United States and Canada between 16,000 to 10,000 years ago, enormous vol- umes of glacial meltwater poured through the Mississippi and Ohio river systems eroding huge volumes of sediment as they flowed. Examination of existing sediment layers shows that the upper Mississippi River eroded and carried away a 12 m (40 ft) thickness of sediments. With removal of this heavy load of sediments, the land rebounded upward, reducing the stresses that held the underlying faults in place.
IndianaIllinois
Missouri
VIII
VIII
VII
VII
IX
IX
X
XI
Kentucky
Tennessee Arkansas
Poplar Bluff
St. Louis
Carbondale Evansville
Paducah
Nashville
Memphis
Little Rock
Mississippi
0
0
200 mi
300 km
Figure 5.36 Map showing estimated Mercalli intensities expected from a recurrence of an 1811–1812 New Madrid earthquake. Intensity VIII and above indicates heavy structural damage. Source: R. M. Hamilton and A. C. Johnston, “Tecumseh’s prophecy: Preparing for the next New Madrid earthquake” in US Geological Survey Circular 1066.
abb22878_ch05_110-142.indd 135abb22878_ch05_110-142.indd 135 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
136 Chapter 5 Earthquakes Throughout the United States and Canada
aftershock from a magnitude 7 or greater earthquake that occurred offshore at a much earlier date.
On 9 November 1727, an earthquake rattled the East Coast from Maine to Del- aware. The epicenter was near Newbury, Massachusetts ( figure 5.40 ), and the shaking caused chimneys and stone walls to fall and cellar walls to collapse. Some uplands were dropped down to become wet lowlands, and some wet lowlands were uplifted and became dry enough to support grasses. Quicksand conditions were common during the earthquake.
Shortly before dawn on a frigid 18 November 1755, the entire eastern seaboard from Nova Scotia to South Carolina was shaken with an earthquake that began offshore from Cape Ann, Massachusetts. In Boston, so many chimneys reportedly toppled that some streets were made impassable by the debris. The seism is estimated to have had a magnitude of about 6.3, but the shaking was so severe that residents reported seeing the land rolling with waves like the surface of the sea. This earthquake occurred just 17 days after the epic earthquakes in Lisbon, Portugal, and it fired up the doom-and-gloom preachers who saw the seism as just pun- ishment for the sins of New Englanders.
Many of these earthquakes may be related to the faults that bound former rift valleys ( figure 5.39 ). The ancient faults may be reactivating and failing due to current stresses. Do all the rift-bounding faults have the potential for future seismic activity? The historic record is not long enough to properly answer this question. But if the answer is yes, then virtually the length of the Atlantic Coastal province could receive a significant shake sometime.
When the next magnitude 6 or greater earthquake strikes the eastern United States, the resultant destruction is likely to be proportionately greater than for a similar seism in the western part of the country. In the East, earthquake energy is transmitted more effectively in the older, more solid rocks, so damages may be experienced over a wider area. Also consider (1) the population density of the East, (2) the large number of older buildings not designed to withstand earth- quake shaking, and (3) the concentration of industrial and power-generating facilities, including nuclear reactors.
ST. LAWRENCE RIVER VALLEY The St. Lawrence is another river whose present path results from occupying an ancient tectonic structure. Some 600 to 500 million years ago, a major rift valley extended through
Intraplate Earthquakes: Eastern North America The large earthquakes of eastern North America share characteristics with those of central North America. Most occur at sites of ancient rift valleys. Most lack significant recent faults. The regions have low strain rates. Figure 5.39 shows some rift arms developed 220 to 180 million years ago as Pangaea was torn apart. Some rift arms succeeded, com- bining to create today’s Atlantic Ocean basin. Other rift arms failed and left behind weakened zones within continents.
NEW ENGLAND New England has a long record of significant earthquakes. On 11 June 1638, just 18 years after the Pilgrims landed in Plymouth, Massachusetts, a sizable earthquake rocked them. It rattled dishes, shook buildings, and in general frightened the Europeans, who were unfamiliar with earthquakes. Due to the limited number of settlements, it is difficult to pinpoint the location of the fault movement that generated this earth- quake. However, a suggested epicentral site lies offshore from Cape Ann; estimates of Mercalli intensity range all the way up to IX, and a magnitude estimate based on felt area is 5.5. It has been suggested that this earthquake was merely an
N
Outline of buried rift complex
Ancient faults
Lower crust
Regional compressive
stress
Outline of buried rift complex
Upper crust
Edge of embayment
Buried rift valley
Ancient magma bodies
Earthquake hypocenters
Missouri Arkansas
Illinois
Indiana
Kentucky Tennessee
Figure 5.37 Schematic block diagram of the Reelfoot rift, the ancient failed rift valley beneath the upper Mississippi River embayment. Large earthquakes are likely caused by present tectonic stresses triggering failures on ancient faults. Source: US Geological Survey Professional Paper 1236L.
abb22878_ch05_110-142.indd 136abb22878_ch05_110-142.indd 136 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Intraplate Earthquakes: Eastern North America 137
the region ( figure 5.40 ). This now-buried rift coincides with most of the significant earthquakes in southeastern Canada. Seisms within the rift valley commonly reach magnitude 7, yet in unrifted continent nearby, the largest earthquakes are usually only in the magnitude 5 range.
The most active area along the St. Lawrence River Valley is an 80 km (50 mi) by 35 km (22 mi) zone near Charlevoix, northeast of Quebec City. Here, earthquakes of magnitudes 6 to 7 occurred in 1534, 1663, 1791, 1860, 1870, and 1925. Why the concentration of large seisms in this one relatively small area? Charlevoix was the site of a meteorite impact some 350 million years ago. The impact caused inten- sive fracturing of the area, including faults in curving pat- terns. These impact-caused fractures are perhaps being reactivated today under the stresses generated by the opening Atlantic Ocean basin.
0
0 200 km
200 mi c. 175 m.y.b.p c. 550 m.y.b.p c. 1,100 m.y.b.p
M M
MC FWLS
SLA
SI A
E C
RC
RT
R RSO
D
Figure 5.38 Map showing approximate locations of buried, ancient rifts in the central United States. Rifting occurred during three principal times—around 220 to 175 million years ago (red); 600 to 500 million years ago (purple); and 1,100 to 1,000 million years ago (green). Some older rifts were apparently rifted again under later plate-tectonic regimes. Rifts are: D, Delaware; EC, East Continent; FW, Fort Wayne; LS, La Salle; MC, Mid- Continent; MM, Mid-Michigan; RC, Rough Creek; RR, Reelfoot Rift; RT, Rome Trough; SIA, Southern Indiana Arm; SLA, St. Louis Arm; and SO, Southern Oklahoma. Source: D. W. Gordon, US Geological Survey Professional Paper 1364.
North America
Nova Scotia rift
Green- land
Grand Banks rift
Connecticut rift
Newark riftMississippi
River
Amazon River
South America
Niger River
Africa
Figure 5.39 Schematic map of rifts that tore at Pangaea about 220 million years ago. Successful rifts combined to open the Atlantic Ocean basin.
CHARLESTON, SOUTH CAROLINA, 1886 Charleston sits alongside a beautiful bay, a charming city with distinguished buildings, wide boulevards, and invit- ing gardens. The presence of the port helped the city develop as a wealthy trading center. Yet Charleston has another side. In the mid-1800s, it was a hotbed of seces- sionist fervor; the first shots of the Civil War were fired over its harbor at Fort Sumter on 12 April 1861. After the war ended four years later, Charleston was
a city of ruins, of desolation, of vacant houses, of widowed women, of rotting wharves, of deserted warehouses, of acres of pitiful and voiceless barrenness.
Yet by the mid-1880s, Charleston had been restored as a cen- ter of wealth, aesthetic buildings, and cultural achievement.
abb22878_ch05_110-142.indd 137abb22878_ch05_110-142.indd 137 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
138 Chapter 5 Earthquakes Throughout the United States and Canada
48
46
44
42
66687072747678
1983 Goodnow
1727 Newbury
1884
1755 Cape Ann5.2mb
5.0Mfa
5.1ML
5.0Mfa 5.5Mfa
1929 Attica
6M
0 300 km
1663 7M 1925 7M
1870 6.5M
1988 Saguenay
St . L
aw re
nc e R
ive r
Quebec City
Montreal Ottawa
1944 Massena 5.9mb
1975 Racquette Lake 3.9mbLg
Albany Boston
New York City
USA
Canada
Boundaries of ancient rift
Figure 5.40 Some earthquake locations in the St. Lawrence River valley area. The approximate location of the 600- to 500-million-year-old rift valley is shown in purple. M fa equals magnitude estimated from the felt area. Large earthquakes northeast of Quebec City lie in the circular Charlevoix seismic zone.
Figure 5.41 Damage from the 1886 earthquake in Charleston, South Carolina. Photo by J. K. Hillers, US Geological Survey.
Even the damages wrought by an 1885 hurricane were not enough to slow the city. Then came 31 August 1886, a typical sultry summer day. At 9:50 p.m., the quiet, breezeless evening was shat- tered by the largest earthquake to occur east of the Appalachian Mountains in historic time. Sixty sec- onds of shaking left 60 people dead, and once again, the remaining citizens had to put their city back together. About 90% of the buildings were damaged or destroyed ( figure 5.41 ).
The earthquake had a magnitude estimated at 7.3. The event produced no surface faulting, so the fault movement may have occurred below 20 km (12 mi) depth. The large magnitude corre- sponds to a rupture length of about 30 km (19 mi) and a rupture width on the fault surface of about 19 km (12 mi). The seism was felt over a large area, and damages were widespread as well ( fig- ure 5.42 ). The large felt area is typical of the eastern United States and is largely attributable to the persistence of longer-period seismic waves (e.g., 1 second), which simply do not die down as quickly as they do in the western United States.
abb22878_ch05_110-142.indd 138abb22878_ch05_110-142.indd 138 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Earthquakes and Volcanism in Hawaii 139
The continental rocks at depth are geologically old and rigid, causing the region to “ring like a bell” and transmit seismic waves far and wide.
How rare is an earthquake of this size for Charleston? Sediments exposed in trench walls, augmented by radio-carbon dates, tell of at least five other similar-sized earthquakes in the area in the past 3,000 to 3,600 years. Thus, large seisms may be expected about every 600 years.
Earthquakes and Volcanism in Hawaii When we think about natural hazards in Hawaii, it is volca- nism that comes to mind. But the movement of magma can cause earthquakes, including large ones ( table 5.5 ). When rock liquefies, its volume expands, and neighboring brittle rock must fracture and move out of the way. The sudden breaks and slips of brittle rock are fault movements that produce earthquakes. When magma is on the move at shallow depths, it commonly generates a nearly continuous swarm of relatively small earthquakes referred to as harmonic tremors.
+ +
+ + + +
+ +
+ +
+
+
+ +
++
++
+
++
+
II–III
II–III
II–III
IV
IV
VI
V
V
VV V
V
VI
IV
VII
VIII
IX
X
A tla
nt ic
O ce
an
0 200
0 200 400 km
400 mi
Figure 5.42 Mercalli intensity map for the 31 August 1886 earthquake near Charleston, South Carolina. Source: G. A. Bollinger, US Geological Survey Professional Paper 1028, 1977.
TABLE 5.5 Some Large Earthquakes in Hawaii
Date Location Intensity Magnitude
2 Apr 1868 Southeast Hawaii X 7.9
5 Oct 1929 Honualoa, Hawaii VII 6.5M s
22 Jan 1938 North of Maui VIII 6.7M s
25 Sep 1941 Mauna Loa, Hawaii VII 6.0M s
22 Apr 1951 Kilauea, Hawaii VII 6.5
21 Aug 1951 Kona, Hawaii IX 6.9
30 Mar 1954 Kalapana, Hawaii VII 6.5
26 Apr 1973 Southeast Hawaii VIII 6.3
29 Nov 1975 Southeast Hawaii VIII 7.2M s
16 Nov 1983 Mauna Loa, Hawaii VII 6.6M s
25 Jun 1989 Kalapana, Hawaii VIII 6.5
15 Oct 2006 Kalaoa, Hawaii VIII 6.7M w
abb22878_ch05_110-142.indd 139abb22878_ch05_110-142.indd 139 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
140 Chapter 5 Earthquakes Throughout the United States and Canada
Kilauea is “supported” on the northwest by the gigantic Mauna Loa volcano and the mass of the Big
Island of Hawaii. However, on its southeastern side, there is less support; Kilauea drops off into the Pacific Ocean. The effects of subsurface magma movement, both com- pressive during injection and extensional during removal, combine with gravitational pull to cause large movements on normal faults.
EARTHQUAKE IN 1975 On 29 November, one of the seaward-inclined normal faults moved suddenly in a 7.2M s seism. It happened at 4:48 a.m., when a large mass slipped for 14 seconds with a movement of about 6 m (20 ft) seaward and 3.5 m (11.5 ft) downward. The movement of this mass into the sea caused tsunami up to 12 m (40 ft) high. Campers sleeping on the beach were rudely awakened by shaking ground; those who didn’t immediately hustle to higher ground were subjected to crashing waves. Two people drowned. This fault movement had an effect on subsurface magma analogous to shaking a bottle of soda pop—gases escaping from magma unleashed an 18-hour eruption featuring magma fountains up to 50 m (165 ft) high.
EARTHQUAKES IN 2006 On 15 October, the Big Island of Hawaii was rocked by two large earthquakes just 7 minutes apart. The first seism was 6.7M w at the hypocentral depth of 40 km (25 mi); the second was 6.0M w at 20 km (12 mi) depth. The earthquakes seem to have resulted from the heavy load that the huge island of Hawaii places on the lithosphere. They occurred where the lithosphere is bent or flexed the most. The initial deeper earthquake resulted from tensional forces pulling rock apart. The following shallower earthquake occurred due to com- pressional forces pushing rock together.
Figure 5.43 shows that the earthquakes below Kilauea Volcano are dominantly near-surface events.
Magma movements also cause larger-scale topographic features and larger earthquakes with magnitudes in the 6 sand 7s. The land surface is commonly uplifted due to the injection of magma below the ground surface. But the land surface is also commonly down-dropped due to withdrawal of magma. Figure 5.44 shows some down-dropped valleys on Kilauea; the valley walls are normal faults.
Mauna Loa NW SE
Kilauea
D ep
th (
km )
Distance (km) 200
60
40
20
0
40 60
Pacific Ocean
SE
Mauna Loa Volcano
Sea level
Ko ae
fa ult
zo ne
H ilin
a fa
ul t
zo ne
NW
Figure 5.44 Schematic block diagrams of the southeastern flank of Kilauea Volcano. Intruding magma (red) forces brittle rock to break and move, generating earthquakes. Gravity-aided sliding down normal faults causes more earthquakes as rock masses slide south-eastward into the ocean and cause rare mega-tsunami.
Figure 5.43 Cross-section showing hypocenters beneath Kilauea Volcano on the flank of the larger Mauna Loa Volcano, southeastern Hawaii, 1970–1983. From F. W. Klein and R. Y. Koyanagi, “The Seismicity and Tectonics of Hawaii” in Geological Society of America, Decade of North American Geology. Vol. N. Copyright © Geological Society of America. Reprinted with permission.
abb22878_ch05_110-142.indd 140abb22878_ch05_110-142.indd 140 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Questions for Review 141
Fault movement is complex; movement lasts only a few seconds at one spot, it speeds up and slows down, it slips different amounts in different sections, and it can trigger other activity in the direction of its movement.
The large left step in the San Andreas fault in the Los Angeles area causes compressive ruptures along east-west- oriented thrust faults, as in the 1971 San Fernando and 1994 Northridge events. More Northridge-type earthquakes are likely.
Prehistoric earthquakes may be interpreted using faulted pond sediments. The amount of offset of sediment layers is proportional to earthquake magnitude. Organic material in sediment layers can be dated by measuring the amount of radioactive carbon present. These techniques were used in 1988 to forecast a 30% probability of a magnitude 6.5 earth- quake in the Loma Prieta area by the year 2018. In 1989, a magnitude 6.9 event occurred.
Southern California may have several large earthquakes in the 21st century. The southern segment of the San Andreas fault is the only one not to have a long rupture in historic time. In prehistory, it has ruptured every 250 years on aver- age, but the last big movement was in 1690. The next big earthquake in California quite possibly will be a magnitude 7.8 event rupturing 300 km (185 mi) of the fault.
Humans have triggered earthquakes by pumping water underground under pressure; by building dams and impound- ing water, which seeps underground under pressure; and by underground explosions of atomic bombs.
Earthquakes occur throughout North America. Most are in the West along the edges of the active plates, but the cen- tral and eastern regions also have earthquakes—not as many, but some of them large.
In the Pacific Northwest, earthquakes occur 30 to 70 km (20 to 45 mi) deep within the subducting oceanic plate as minerals change form due to increased temperature and pres- sure. At the surface, strike-slip faults rupture the ground, as in Seattle.
The Basin and Range province between eastern California and central Utah is an actively extending area. For example, Nevada has about doubled in west-east width in the past 30-million years. Normal faults accommodate most of the extension, unleashing earthquakes up to magnitude 7.3.
In the central United States and eastern North America, ancient rift valleys remain from failed spreading centers. The ancient rifts today are zones of weakness whose faults can be reactivated due to long-distance effects of Atlantic plate spreading and Pacific plate collision. The Reelfoot rift, occupied today by the Mississippi River, had earthquakes in 1811 and 1812 with moment magnitudes (M w ) of 7.3, 7.0, 7.0, and 7.5. Other rift valleys are associated with earthquakes throughout North America.
Charlevoix, Quebec, has frequent earthquakes up to magnitude 7. The region was intensely fractured by the
Summary impact of an ancient asteroid, and the fractured rocks appar- ently move due to stresses within the moving plates.
The underground movement of magma in Hawaii gener- ates earthquakes up to magnitude 7.9 by forcefully rupturing brittle rocks. Land is uplifted as magma is injected and dropped down when magma is removed; these land move- ments may be sudden, earthquake-generating events, and some create tsunami as well.
Terms to Remember avalanche 115 blind thrust 113 coal 112 directivity 113 earthquake weather 119 embayment 132 failed rift 135 fracking 120 friction 111
harmonic tremors 139 hydraulic fracturing 120 neotectonics 116 paleoseismology 116 photosynthesis 117 quicksand 136 scarp 130 thrust fault 113
Questions for Review 1. Sketch a map, and explain the elastic-rebound theory of
faulting. How has the theory been modified in recent years? 2. Draw a cross-section of a blind-thrust fault, such as the one
that affected Northridge in 1994. Why was ground shaking so intense in this earthquake?
3. Draw a cross-section, and explain how faulted pond sediments can be used to tell the magnitudes and frequencies of ancient earthquakes.
4. During an earthquake, you probably will be safest if you ________, ________, and ________.
5. Explain three ways that humans have caused or triggered earthquakes.
6. Earthquakes are most abundant in which two U.S. states? In which Canadian province?
7. What types of evidence indicate major movements on surface faults in Washington?
8. As the depth to a hypocenter increases, how does that affect surface shaking?
9. What tectonic process has affected the Basin and Range province, from eastern California to Utah, during the past 30 million years? How much stretching has occurred across it? What is the orientation of the ranges and basins, and what does this tell us about the direction of stretching?
10. What type of fault movement best characterizes the Basin and Range province? What are the highest magnitude earthquakes generated there in the past century?
11. “Stable” central and eastern North America have earthquakes clustered in distinct areas. What is a likely control on these earthquake locations?
12. When and how did the Reelfoot rift form? Explain its history of earthquakes. What is the name of the strongest historical earthquake swarm that occurred here? What does the future hold?
abb22878_ch05_110-142.indd 141abb22878_ch05_110-142.indd 141 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
142 Chapter 5 Earthquakes Throughout the United States and Canada
13. How does intrusion of magma on the flanks of Kilauea volcano on Hawaii generate earthquakes?
14. What are the harmonic tremors experienced in Hawaii? What do they tell us?
15. Why are the western United States and Canada much more seismically active than the central and eastern regions? Does the smaller number of earthquakes in the central and eastern United States and Canada mean that we don’t have to worry about a “big one” there?
16. Besides an actual downward subduction movement, how can a subducting plate generate earthquakes, as in Washington?
17. What is hydraulic fracturing? How can it trigger an earthquake?
18. What is meant by directivity of seismic waves? How significant is directivity?
19. Are seismic waves amplified more in hard rocks or in soft sediments? What happens to amplitude of seismic waves when they enter soft sediments? (See discussion under ShakeMaps.)
20. What event, unrelated to plate tectonics, weakened the crust along part of the St. Lawrence River Valley, making it easier to reactivate faults there?
Questions for Further Thought 1. So-called psychics are commonly quoted in the media
predicting that California will break off along the San Andreas fault and sink beneath the sea. Is this possible? Why not?
2. Some people suggest that earthquakes usually occur at certain times of day. Does this make sense? Is there a pattern to the times of the earthquakes discussed in this text?
3. Assess the earthquake hazard in Salt Lake City. 4. What controls the course of the Rio Grande in New Mexico?
Is there an earthquake threat also? 5. Compare the ability to withstand earthquake shaking of
downtown buildings and bridges in West Coast cities to that of structures in the mid-continent or East Coast of the United States.
6. Humans can trigger earthquakes, but should we? Can we set off medium-size earthquakes in a controlled fashion that will prevent a large earthquake in an area? Make a list of pros and cons for earthquake control.
7. Could the Chinese earthquake of 2008 have been triggered by the recent filling of a nearby reservoir?
8. What are possible causes of the 5.8M W earthquake in Virginia on 23 August 2011?
abb22878_ch05_110-142.indd 142abb22878_ch05_110-142.indd 142 27/12/12 3:50 PM27/12/12 3:50 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanic Eruptions: Plate Tectonics and Magmas 6
“The simplest explanation that covers all the facts is the best one.”
—Occam’s Razor , attributed to William of Occam, c. 1295–1349
CHAPTER
LEARNING OUTCOMES Volcanic eruptions can be overwhelming events. Understanding of magma types and plate-tectonic settings explains a lot of volcanic behaviors. After studying this chapter, you should:
• know the plate-tectonic settings of volcanoes.
• understand the variations in magma characteristics that control peaceful versus catastrophic eruptions.
• comprehend the roles of magma gas content and pressure reductions in volcanic eruptions.
• be able to explain the three Vs of volcanism, and relate them to volcanic eruption styles, and to volcanic landforms.
• be familiar with the Volcanic Explosivity Index.
• be able to describe hot spots.
OUTLINE • How We Understand Volcanic Eruptions
• Plate-Tectonic Setting of Volcanoes
• Chemical Composition of Magmas
• Viscosity, Temperature, and Water Content of Magmas
• How a Volcano Erupts
• The Three Vs of Volcanology: Viscosity, Volatiles, Volume
Lava meets the sea on Hawaii. Photo from © StockTrek/Getty Images RF.
In te
r n
al E
n er
g y
abb22878_ch06_143-173.indd 143abb22878_ch06_143-173.indd 143 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
144 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Volcanoes deal out overwhelming doses of energy no human can survive. The dangers of volcanic eruption are obvious, but the quiet spells between active volcanism are seductive. Some people are lured to volca- noes like moths to a flame, even those who should know better. On 14 January 1993, volcanologists attending a workshop in Colombia, as part of the international decade of natural disaster reduction, hiked into the sum- mit crater of Galeras Volcano to sample gases and mea- sure gravity. They were looking for ways to predict imminent eruptions. The volcano had been quiet since July 1992, but during their visit, an unexpected, gas- powered secondary eruption killed six in the scientific party—four Colombians, a Russian, and an Englishman. Their deaths were not an unusual event ( table 6.1 ). They serve as a small-scale example of the larger drama played out when a volcano suddenly buries an entire city. During long periods of volcanic quiescence, people tend to build cities near volcanoes. For example, 400,000 people live on the flanks of Galeras Volcano, defying the inevitability of a large, life-snuffing eruption.
An individual volcano may be active for millions of years, but its eruptive phases are commonly separated by centuries of inactivity, lulling some into a false sense of security. Around 410 BCE , Thucydides wrote, “History repeats itself.” We know well that those who do not learn the lessons of history are doomed to repeat them. Every year, people inad- vertently sacrifice their lives to volcanic eruptions.
How We Understand Volcanic Eruptions Two primary building blocks of knowledge are paramount to understanding volcanic eruptions:
1. Plate tectonics has given us great insight into earth- quakes; now it will help us understand volcanoes.
2. Magmas (liquid rocks) vary in their chemical composi- tion, their ability to flow easily, their gas content, and their volume. These variations govern whether eruptions are peaceful or explosive.
In this chapter we first take a brief look at plate tectonics and volcanism. Then we examine magma variations and how they control eruptive style. Finally, we apply this knowledge globally to understand why volcanoes occur where they do, why only some volcanoes explode violently, and how volca- noes can kill.
Plate-Tectonic Setting of Volcanoes Convection of heat in the mantle drives plate tectonics. More than 90% of volcanism is associated with the edges of tectonic plates ( figure 6.1 ). Most other volcanism occurs above hot spots (see figure 2.23). More than 80% of Earth’s magma extruded through volcanism takes place at the oce- anic spreading centers. Solid, but hot and ductile, mantle rock rises upward by convection into regions of lower pres- sure, where up to 30–40% of the rock can melt and flow easily as magma on the surface ( figure 6.2 ). The worldwide rifting process releases enough magma to create 20 km 3
(about 5 mi 3 ) of new oceanic crust each year. Virtually all this volcanic activity takes place below sea level and is thus difficult to view.
Subduction zones cause the tall and beautiful volcanic mountains we see at the edges of the continents, but the volume of magma released at subduction zones is small com- pared to that of spreading centers. Subduction zones account for the eruption of 7–13% of all magma. The down-going plate carries oceanic-plate rock covered with water-saturated sediments into much hotter zones ( figure 6.2 ). The presence of water lowers the melting point of rock. Rising magma
TABLE 6.1 Volcanologists Killed by Eruptions
Year Volcano Total
Deaths Dead
Volcanologists 1951 Kelut, Indonesia 7 3
1952 Myojin-sho, Japan 31 9
1979 Karkar, New Guinea 2 2
1980 St. Helens, United States 62 2
1991 Unzen, Japan 44 3
1991 Lokon-Umpong, Indonesia 1 1
1993 Galeras, Colombia 9 6
1993 Guagua Pichincha, Ecuador 2 2
2000 Semeru, Indonesia 2 2
abb22878_ch06_143-173.indd 144abb22878_ch06_143-173.indd 144 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Plate-Tectonic Setting of Volcanoes 145
Basaltic ocean ic crust with sediments on
top Sea level
Melting of mantle wedge above subducting plate aided by water from sediments
Folded sediments
Andesitic-rhyolitic volcanoes
Continental crust
Partial melting of continental crust
Subduction zone Spreading center
Basaltic volcanism
Rising of partially melted asthenosphere
Asthenosphere
Deep-ocean trench
Figure 6.2 Idealized cross-section showing production of basaltic magma at spreading centers. Plates pull apart, and some asthenosphere liquefies and rises to fill the gap. Andesitic-rhyolitic magmas are created above subduction zones, where rising magma partially melts the continental crust on its way up, thus altering the melt by increasing SiO 2 content and viscosity.
Explosive eruptions Numerous volcanoes
Andesitic-rhyolitic magma
Subduction zone
Spreading center
Tr an
sf o
rm f
au lt
Li ttl
e or
n o
vo lc
an is
m Tran sfo
rm fau
lt Little or no volcanism
Voluminous basaltic magma
Volcanic mountain ranges Peaceful eruptions
Figure 6.1 An idealized oceanic plate showing styles of volcanism.
partially melts some of the continental crust it passes through. This adds new melt of different composition to rising plumes of magma. Each rising plume has its own unique chemical composition.
Transform faults and continent-continent collision zones have little or no associated volcanism. Thinking three- dimensionally, this is understandable. At a transform fault, the two plates simply slide past each other in a horizontal sense and at all times keep a quite effective “lid” on the hot asthenosphere some 100 km (60 mi) below. At continent- continent collisions, the continental rocks stack up into extra-thick masses that deeply bury the hot mantle rock, making it difficult for magma to rise to the surface.
From a volcanic disaster perspective, the differences are clear. Oceanic volcanoes are relatively peaceful, whereas subduction-zone volcanoes are explosive and dangerous. Ironically, humans tend to congregate at the seaward edges of the continents, where the most dangerous volcanoes operate.
People commonly speculate upon whether an individual volcano is active, dormant, or extinct. Because of the strong hope that a volcano is extinct and the nearby land is thus available for use, many dormant volcanoes are misclassified. But consider this: a subduction zone commonly lasts for tens of millions of years, and its province of volcanoes is active for the entire time. An individual volcano may be active for hundreds of thousands to several million years, despite “slumbers” of centuries between eruptions. As a general rule, if a volcano has a well-formed and aesthetic conical shape, it is active. A pretty shape is dangerous.
abb22878_ch06_143-173.indd 145abb22878_ch06_143-173.indd 145 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
146 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
A Classic Disaster Eruption of Mount Vesuvius, 79 ce The most famous of all volcanoes is probably Vesuvius in Italy, and the most famous of all its eruptions must be those of 79 CE . It was then that the cities of Pompeii and Herculaneum were buried and forgotten for more than 1,500 years. A warning of the natural hazards near Vesuvius arrived on 5 February 62 CE , when a major earthquake destroyed much of Pompeii and caused serious damage in Herculaneum and Neapolis (Naples). Additional earthquakes, although not as large as this first one, were a common occurrence for the next 17 years.
Pompeii had been a center of commerce for centuries. In 79 CE , the city had a population of about 20,000 people, 8,000 of whom were slaves. Robert Etienne described it as:
An average city inhabited by average people, Pompeii would have achieved a comfortable mediocrity and passed peace- fully into the silence of history, had the sudden catastrophe of the volcanic eruption not wiped it from the world of the living.
The 24th of August 79 CE was a warm summer day, but then Vesuvius began erupting and the day became even hotter. Vesu- vius blew out 4 km 3 (1 mi 3 ) of volcanic material. About half of the old volcanic cone was destroyed. A modern example of a similar eruption occurred in 1991 at Mount Pinatubo in the Philippines ( figure 6.3 ). In Pompeii, great clouds of hot gas and volcanic ash flowed across the city, killing the people who had not fled ( figure 6.4 ). Today, the excavated city is a major attrac- tion for tourists ( figure 6.5 ).
In 79 CE , the fine volcanic ashes settling out from the great heights of the eruption cloud affected a large region. Pliny the Younger was at Misenum and wrote:
And now came the ashes, but at first sparsely. I turned around. Behind us, an ominous thick smoke, spreading over the earth like a flood, followed us. “Let’s go into the fields while we can still see the way,” I told my mother—for I was afraid that we might be crushed by the mob on the road
Figure 6.3 The first big explosive blast from Mount Pinatubo occurred on 15 June 1991. Photo by R. S. Culbreth, US Geological Survey .
Figure 6.4 Body cast of a man killed by a flow of hot gas and volcanic ash from Vesuvius in late August in the year 79 CE .
Figure 6.5 Tourists walk through the heart of Pompeii away from Vesuvius, the volcano that destroyed the city in 79 CE . Photo by Pat Abbott.
in the midst of darkness. We had scarcely agreed when we were enveloped in night—not a moonless night or one dimmed by cloud, but the darkness of a sealed room with- out lights. To be heard were only the shrill cries of women, the wailing of children, the shouting of men. Some were calling to their parents, others to their children, others to their wives—knowing one another only by voice. Some wept for themselves, others for their relations. There were those who, in their very fear of death, invoked it. Many lifted up their hands to the gods, but a great number believed there were no gods, and that this was to be the world’s last eternal night.
abb22878_ch06_143-173.indd 146abb22878_ch06_143-173.indd 146 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Plate-Tectonic Setting of Volcanoes 147
In the coastal city of Herculaneum, 300 skeletons were found in lifelike positions in boat chambers at the beach. The skeletons of these people killed by the eruption testify to the lethal energy they experienced. The people had not been battered or suffo- cated; they did not display any voluntary self-protection reactions or agony contortions. In other words, their vital organs stopped functioning in less than a second, in less time than they could consciously react. The types of bone fractures, tooth cracks, and bone coloration indicate the victims were covered by volcanic material at about 500°C (930°F). At this temperature, their soft tissues vaporized; their feet flexed in an instantaneous muscle contraction ( figure 6.6 ).
Figure 6.6 Feet of a child killed by an eruption from Vesuvius in 79 CE . The flexed toes and feet were an involuntary contraction when surrounded by 500°C (930°F) volcanic material. Death was instantaneous. Photo from Mastrolorenzo, G., Petrone, P., Pagano, M., Incoronato, A., Baxter, P., Canzanella, A., Fattore, L., in Nature 410:769, 2001.
Figure 6.7 Mount Vesuvius, near center of photo, is sur- rounded by the urbanized Naples region housing 3 million people. Photo from NASA.
The timing of the major eruptions of Vesuvius offers an inter- esting lesson. Apparently Vesuvius did not have a major eruption from the 7th century BCE until 79 CE . People had at least 700 years to lose their fears and yield to the allure of the rich agricultural soils on Vesuvius. After 79 CE , large eruptions occurred more often: in the years 203, 472 (ash blown over much of Europe), 512, 685, 993, 1036 (first lava flows in historic time), 1049, and 1138–1139. Then nearly 500 years passed—plenty of time to forget the past and recolonize the mountain. But in 1631, Vesuvius poured out large volumes of lava that destroyed six towns; mudflows ruined another nine towns, and about 4,000 people perished. The two long periods of volcanic quiescence in the last 2,700 years seem like long times to short-living, land-hungry humans, but this is the time schedule of an active volcano. Humans’ lack of appreciation for the time involved between eruptions leads them to falsely regard many active volcanoes as extinct.
Since 1631, the eruptions from Vesuvius have been smaller and not as dangerous. There were 18 eruption cycles between 1631 and 1944; each lasted from 2 to 37 years, with quiet intervals rang- ing from 0.5 to 6.8 years. Since 1944, Vesuvius has been quiet. Is this interval of calm setting the stage for another major eruption? We do not know for sure, but almost 3 million people live within reach of Vesuvius today, including about 1 million on the slopes of the volcano ( figure 6.7 ).
A Classic Disaster
abb22878_ch06_143-173.indd 147abb22878_ch06_143-173.indd 147 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
148 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Why do spreading-center volcanoes have relatively peaceful eruptions? And why do subduction-zone volcanoes explode violently? The answers to these questions are found in knowing how different magmas behave.
Chemical Composition of Magmas Although there are 92 naturally occurring elements, a mere eight make up more than 98% of the Earth’s crust ( table 6.2 ). The next four most abundant elements add another 1.2% to the crust, bringing the weight percent contributed by these 12 elements to 99.23%. The remaining 0.77% includes gold, silver, copper, carbon, sulfur, tin, and many other familiar elements.
Oxygen and silicon are so abundant that their percent- ages dwarf those of all other elements. Oxygen atoms carry negative charges (–2), while silicon atoms are positively charged (+4). As magma begins cooling, some silicon and oxygen atoms bond. Silicon and oxygen link up with four oxygen atoms (4 × –2 = –8) surrounding a central silicon atom (+4) to form the silicon-oxygen tetrahedron (SiO 4 ) ( figure 6.8 ). The SiO 4 tetrahedron presents a –4 charge on its exterior that attracts and ties up positively charged atoms. After negatively charged oxygen, the 11 elements of greatest abundance are all positively charged ( table 6.2 ); they are attracted to, and bound up by, oxygen. This process is so common that elemental abundances in the crust are usually
O
O
O
OSi
Figure 6.8 A silicon atom with +4 charge is linked to four oxygen atoms, each with a –2 charge.
TABLE 6.2 Common Elements of the Earth’s Crust (weight %)
Eight Most Common Oxygen(O �2 ) 45.20%
Silicon (Si �4 ) 27.20
Aluminum (Al �3 ) 8.00
Iron (Fe �2,�3 ) 5.80
Calcium (Ca �2 ) 5.06
Magnesium (Mg �2 ) 2.77
Sodium (Na �1 ) 2.32
Potassium (K �1 ) 1.68
Total 98.03%
Next Four Most Common Titanium (Ti �3,�4 ) 0.86%
Hydrogen (H �1 ) 0.14
Phosphorus (P �5 ) 0.10
Manganese (Mn �2,�3,�4 ) 0.10
Total 99.23%
listed in combination with oxygen (as oxides). The weight percentages of elements are quite different for continental versus oceanic crust ( table 6.3 ). Marked differences in sili- con dioxide percentages produce magmas of variable erup- tive behavior.
Viscosity, Temperature, and Water Content of Magmas Liquids flow freely; their volumes are fixed, but their shapes can change. Liquids vary in how they flow; some flow quickly, some flow slowly, and some barely flow at all. The fluidity of a liquid is measured by its viscosity, its internal resistance to flow; viscosity may be thought of as a measure of fluid friction. The lower the viscosity, the more fluid is the behavior. For example, tilt a glass of water and watch it flow quickly; water has low viscosity. Now tilt the same glass filled with honey and watch the slower flow; honey has higher viscosity. Low-viscosity magma flows somewhat like ice cream on a hot day. High-viscosity magma barely flows.
TABLE 6.3 Crustal Elements in Weight-Percent Oxides
Continental Crust Oceanic Crust SiO 2 60.2% SiO 2 48.7%
Al 2 O 3 15.2 Al 2 O 3 16.5
Fe 2 O 3 2.5 Fe 2 O 3 2.3
FeO 3.8 FeO 6.2
CaO 5.5 CaO 12.3
MgO 3.1 MgO 6.8
Na 2 O 3.0 Na 2 O 2.6
K 2 O 2.9 K 2 O 0.4
abb22878_ch06_143-173.indd 148abb22878_ch06_143-173.indd 148 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Viscosity, Temperature, and Water Content of Magmas 149
In Greater Depth Minerals and Volcanic Rocks The eight most common elements bond in different configurations to make up hundreds of different minerals. The process of mineral formation in a cooling magma is called crystallization. Just as with elements, a degree of simplicity occurs in a crystallizing magma because the overwhelming majority of Earth’s crust is composed of just eight common rock-forming minerals. Laboratory experiments and microscopic examination of rock-forming minerals have shown the order in which these minerals crystallize from a cooling magma ( figure 6.9 ).
Magmas at the surface with temperatures of around 800° to 1,300°C (1,470° to 2,370°F) have two separate lines of mineral growth:
1. Iron and magnesium link up with the silicon-oxygen tetrahe- dron as magma temperature decreases to sequentially form four distinct and discontinuous families of minerals—olivine, pyroxene, amphibole, and biotite mica.
2. Calcium combines with Al and SiO 4 to begin forming the pla- gioclase feldspar family, a continuous and gradational series of minerals. As temperature decreases, progressively more sodium (and less calcium) is locked within the plagioclase crys- tal structure. By the time magma has cooled down to the 800° to 1,000°C (1,470° to 1,830°F) range, it is largely depleted in Fe, Mg, and Ca. Now potassium crystallizes within muscovite mica and potassium-rich feldspar minerals, and excess Si and O combine without other elements to make the mineral quartz.
Rock types
Basalt Calcium rich
Sodiu m ric
h
Potassium-rich feldspar Muscovite mica
C ontinuous series of
P lagioclase feldspar m
inerals
Quartz
Olivine
Pyroxene
Amphibole
D is
co nt
in uo
us s
er ie
s of
ir on
-
an d
m ag
ne si
um -r
ic h
m in
er al
s
Higher temperature
Earliest crystals
Lower temperature
Latest crystals
Biotite mica
Andesite
Rhyolite
G radational solid-solution series
Figure 6.9 Order of crystallization of minerals from a magma cooling at depth.
TABLE 6.4 Igneous Rock Types
Magma Type Plutonic Rock Volcanic Rock SiO 2 < 55% Gabbro Basalt
SiO 2 = 55–65% Diorite Andesite
SiO 2 > 65% Granite Rhyolite
Just as elements combine to make minerals, so minerals aggre- gate to make rocks (see page 5). Magmas have a broad range of compositions, resulting in many different types of igneous rocks that generations of geologists have classified into a dizzying array of rock names. Nonetheless, a working understanding can be gained by considering only three magma types and the three clans of igneous rocks that form from them. The rock types are based on their silicon and oxygen (SiO 2 ) percentages ( table 6.4 ). If the magma cools and solidifies below the surface, it crystallizes as plutonic rock, named for Pluto, the Roman god of the underworld. If the magma reaches the surface, it forms volcanic rock, named for Vulcan, the Roman god of fire. The left side of figure 6.9 shows three main types of volcanic rock next to their respective mineral compositions. Table 6.4 tells more about the rocks, and figure 6.10 shows pictures of the rocks. Their magmas behave differently due to their varying temperatures, water contents, and viscosities.
abb22878_ch06_143-173.indd 149abb22878_ch06_143-173.indd 149 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
150 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
commonly does not flow. Table 6.5 also states that about 80% of the magma reaching Earth’s surface is basaltic, with only about 10% andesitic and 10% rhyolitic. Why the difference? Basaltic magma is produced in great abundance by partial melting of the mantle. The lower viscosity of basaltic magma helps it reach the surface, especially at spreading centers and other oceanic settings. Much basaltic magma is also produced at subduction zones, but as it rises through continents, its composition changes as it incorporates continental rock with its high SiO 2 content. During the process of rising, the magma compositions become more andesitic or rhyolitic. The more viscous rhyolitic magmas are so sluggish that they tend to be trapped deep below the surface where they cool, solidify, and grow into the larger mineral crystals of plutonic rocks, such as granite.
In magma, water is the most abundant dissolved gas. As magma rises toward the surface and pressure decreases, water dissolved in the hot magma becomes gas and forms steam bubbles. Basaltic magma is low in dissolved water content, helping make eruptions peaceful. Rhyolitic magma has dissolved water contents up to 6%; as it rises and steam bubbles form, they have difficulty escaping from the high- viscosity magma and have to burst their way out.
When the basaltic volcanoes of Hawaii and Iceland begin to erupt, it is a tourist event. Although such an eruption makes a thrilling show, it is a relatively peaceful happening. Why is it safe? Because it does not contain much dissolved water, and the dissolved gases escape from the low-viscosity magma with relative ease ( figure 6.11 ). Compare this behav- ior to the eruption of a rhyolitic magma of lower temperature, greater dissolved water content, higher percentage of SiO 2 , and very high viscosity. When rhyolitic magma oozes out
The viscosity of magma is changed by various means:
1. Higher temperature lowers viscosity; it causes atoms to spread farther apart and vibrate more vigorously. Thus atomic bonds break and deform more, resulting in increased fluidity. Consider the great effect of tempera- ture on magma ( table 6.5 ). At 600°C (1,100°F), magma viscosity is five orders of magnitude (100,000 times) more viscous than at 900°C (1,650°F).
2. Silicon and oxygen (SiO 2 ) increase the viscosity of magma because they form abundant silicon-oxygen tet- rahedra (see figure 6.8 ) that link up in chains, sheets, and networks, creating more joins and bonds between atoms, which in turn make flow more difficult.
3. Increasing content of mineral crystals increases viscos- ity. Magma is a mixture of liquid and the minerals that have crystallized from it. The mineral content of magma varies from none to the majority of the mass.
Magma contains dissolved gases held as volatiles; their solubility increases as pressure increases and as temperature decreases. You can visualize the pressure–temperature rela- tions using a bottle of soda pop. Carbon dioxide (CO 2 ) gas is dissolved in the soda pop and kept under pressure by the bottle cap. Pop the cap off the bottle, reducing pressure, and some volatiles escape. As the uncapped bottle warms, more volatiles are lost.
A good grasp of volcanic behavior can be gained by con- sidering the properties of three types of magma ( table 6.5 ). Notice that the highest temperatures and lowest SiO 2 contents are in basaltic magma, giving it the lowest viscosity and easi- est fluid flow. The lowest temperatures and highest SiO 2 contents occur in rhyolitic magma, material so viscous that it
(a) (b) (c)
Figure 6.10 Volcanic rock types. (a) Basalt is dark-colored and finely crystalline. Bust of Livia Drusilla, wife of Augustus Caesar, ~31 BCE . 32 cm tall, Louvre Museum, Paris. (b) Andesite is medium-colored volcanic rock with plagioclase feldspar minerals and, in this sample, fragments of other volcanic rocks. Santiago Peak Volcanics, San Diego. (c) This sample of rhyolite contains quartz and feldspar minerals in great abundance. Poway, California. Photos (b) and (c) by Pat Abbott.
In Greater Depth (Continued)
abb22878_ch06_143-173.indd 150abb22878_ch06_143-173.indd 150 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Viscosity, Temperature, and Water Content of Magmas 151
onto the ground surface, the pressure within the magma is reduced and the dissolved gases expand in volume. But how do gases escape from their entrapment in sticky magma? By exploding ( figure 6.12 ). Spectators at the eruption of rhyo- litic magma frequently die. When it comes to volcanic haz- ards, the greatest problem is how much gas is in the magma and how easily the dissolved gases can escape from the magma. As Frank Perret stated: “Gas is the active agent, and magma is its vehicle.”
TABLE 6.5 Comparison of Three Types of Magma
Basaltic Andesitic Rhyolitic Volume at Earth’s Surface 80% 10% 10%
SiO 2 Content 45–55% 55–65% 65–75%
Increasing SiO 2
Temperature of Magma 1,000–1,300°C 800–1,000°C 600–900°C
Decreasing temperature
Viscosity Low (melted ice cream) High (toothpaste)
Increasing viscosity
Water Dissolved in Magma ~0.1−1 weight % ~2−3 weight % ~4−6 weight %
Increasing water
Gas Escape from Magma Easy Difficult
Increasing difficulty
Eruptive Style Peaceful Explosive
Increasing explosiveness
Rock Description Black to dark gray; contains Ca-plagioclase, pyroxene, olivine
Medium to dark gray; contains amphibole, pyroxene, intermediate Ca-Na-plagioclase
Light-colored; contains quartz, K-feldspar, biotite, Na-plagioclase
Figure 6.11 Peaceful eruption of low-viscosity magma with easy separation of volatiles (gas) from magma, Surtsey, Iceland. Photo courtesy of Pat Abbott.
Figure 6.12 Eruption from Paricutin Volcano, Michoacan, Mexico. Photo from US Geological Survey .
PLATETECTONIC SETTING OF VOLCANOES REVISITED Knowing about magma viscosity and volatile content allows us to revisit our earlier questions about plate tectonics and volcanism. Why does the vast majority of Earth’s magma pour out at spreading centers and in relatively peaceful erup- tions? And why does the magma above subduction zones commonly explode violently? Spreading centers operate in
abb22878_ch06_143-173.indd 151abb22878_ch06_143-173.indd 151 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
152 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
oceanic crust, and subducting plates commonly are pulled beneath continental crust. The chemistries of oceanic crust and continental crust are different (see table 6.3 ), their mag- mas are different (see table 6.5 ), and their volcanic behav- iors differ.
Spreading centers are ideal locations for volcanism because (1) they sit above the high-temperature astheno- sphere, (2) the asthenosphere rock has low percentages of SiO 2 , and (3) the oceanic plates pull apart, causing hot asthe- nosphere rock to rise, experience lower pressure, and change to magma that continues to rise. This magma is high- temperature, low SiO 2 , low-volatile content, low-viscosity basalt, allowing easy escape of gases (see figure 6.2 ). Spread- ing centers combine all the factors that promote the peaceful eruption of magma.
When a subducting oceanic plate reaches a depth of about 100 km (over 60 mi), magma is generated and rises toward the surface (see figure 6.2 ). The subducting plate stirs up the mantle, causing the hotter rock at depth to rise and then melt as pressure decreases. A significant reason magma forms here is that the subducting plate carries a cover of sediments, water, and hydrated minerals down with it. Water, even in slight amounts, promotes partial melting by lowering the temperature necessary for rock to melt. The partial melt- ing process affects only those minerals with lower melting temperatures. As this partial melt rises upward, it in turn melts part of the overlying crust to produce magmas of highly variable compositions ( figure 6.13 ). Magma compositions depend on the amount of crustal rock melted and incorpo- rated into the rising magma. In general, in the subduction- zone setting, magma temperature decreases while SiO 2 , water content, and viscosity increase. All these changes in magma add to its explosive potential.
How a Volcano Erupts Earth’s internal energy flows outward as heat (see chapter 2). The eruptions of volcanoes are rapid means for Earth to expel some of its internal heat.
A volcanic eruption begins with heat at depth. Super- heated rock will rise to levels with lower pressure, and some solid rock may change phase to liquid magma, resulting in volume expansion and leading step-by-step to eruption.
Magma is generated by the melting of existing rock. Rock may melt by (1) lowering the pressure on it, (2) raising its temperature, or (3) increasing its water content. How do most rocks melt? The two most relevant melting agents are reductions in pressure (decompression) and increases in vol- atile content (mostly water).
Most magma is generated by decreasing the pressure on hot rock. For example, as the solid, but mobile, hot rock of the mantle rises upward, it experiences progressively less pres- sure and spontaneously melts, without the addition of more heat. Melting caused simply by a decrease in pressure is called decompression melting. The process of decompression
Dominantly andesitic volcanism
High-level magma chambers
Andesite production
Crust
Mantle
Partial melting of crust begins
Basaltic dike injection
Ponding of basalt – 70 km
– 35 km
Basaltic flux from depth
Figure 6.13 Schematic cross-section of magma rising from a subduction zone and being contaminated by crustal rocks en route.
melting is so important that it is worth restating: most of the rock that melts to form magma does so because the pressure on it decreases, not because more heat is added.
The largest nearby reservoir of superhot, ready-to-melt rock exists in the nearly molten asthenosphere. This rock, hot enough to flow without being liquid, is the main source of magma. As this superheated rock rises, the pressure on it decreases, allowing some rock to melt. The hot, rising rock- magma mixture also raises the temperature of rock it passes through, thus melting portions of the overlying rocks.
If pressure in the asthenosphere or lithosphere is decreased, some rock melts, with a resultant increase in vol- ume that causes overlying rocks to fracture. The fractures allow more material to rise to lower pressure levels, causing more rock to liquefy. For example, at a depth of 32 km (20 mi), basaltic rock melts at 1,430°C (2,600°F), but this same rock will melt at only 1,250°C (2,280°F) at the Earth’s surface. Since upward-moving rock/magma reaches ever-lower pres- sures, rising rock can liquefy and magma can increase in fluidity, which in turn causes more superheated rock to become magma.
Magma at depth does not contain gas bubbles because the high pressure at depth keeps volatiles dissolved in solution. But as magma rises toward the surface, pressure continually decreases, and gases begin to come out of solution, forming bubbles that expand with decreasing pressure ( figure 6.14 ). The added lift of the growing volume of gas bubbles helps propel magma upward through fractures or pipes toward an eruption. Gas bubbles continue increasing in number and volume as magma keeps rising upward to lower pressures. As gas-bubble volume increases, the gas can overwhelm magma, fragmenting the magma into pieces that are carried up and out by a powerful gas jet ( figure 6.15 ). Upon escape from the volcano, the gas jet draws in air, which adds to buoyancy in the turbulent, rising plume (see figure 6.14 ).
abb22878_ch06_143-173.indd 152abb22878_ch06_143-173.indd 152 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
How a Volcano Erupts 153
with a low concentration of water only leads to slow flows or no flow as the magma oozes upward and builds a dome.
The most important requirement for explosive eruptions is high concentrations of volatiles (mainly water). Volatiles drive explosive eruptions. Given a high concentration of water, even a basalt magma can erupt violently, as occurred at Hawaii in 1790. Rhyolitic magma is often associated with explosive eruptions because of its high content of water (see table 6.5 ). Water concentration in magma plays a controlling role, and viscosity plays a secondary role, in determining the peaceful versus explosive style of eruption.
Different volcanic behaviors have been classified accord- ing to the eruptive style of individual volcanoes ( figure 6.16 ).
Magma chamber
Rising magma
(expanding gas bubbles)
Gas jet with magma pieces
(dissolved gases)
Magma fragmentation
Air Buoyant plume Air
Figure 6.14 Anatomy of an eruption. As magma rises to levels of lower pressure, gas comes out of solution, forming bubbles that overwhelm the magma and create a gas jet leading to a buoyant plume.
Figure 6.15 Remarkable view into the crater of Mount Pinatubo just as a major explosive eruption was beginning its upward blast, 1 August 1991. Photo from NOAA.
ERUPTION STYLES AND THE ROLE OF WATER CONTENT Whether a volcanic eruption is peaceful or explosive depends significantly on the concentration of water in the magma. For example, if all magmas contained low concentrations of water (such as 0.3 weight %; see table 6.5 ), there would be no highly explosive eruptions. Even a high-viscosity rhyolite magma
Hawaiian type
Icelandic type
Eruption Type
Strombolian type
Vulcanian type
Plinian type
High water content, high viscosity (andesite to rhyolite)
Low water content, low viscosity (basalt)
Low water content, low viscosity (basalt)
Moderate water content, low to moderate viscosity (basalt to andesite)
Moderate to high water content, moderate to high viscosity (basalt to rhyolite)
Composition
Figure 6.16 Some types of volcanic eruptions.
abb22878_ch06_143-173.indd 153abb22878_ch06_143-173.indd 153 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
154 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Nonexplosive eruptions are commonly subdivided into Ice- landic and Hawaiian types. Strombolian types are somewhat explosive. Explosive eruptions can be described as Vulcanian or Plinian types. These classifications are just for general pur- poses; each volcano varies in its eruptive behavior over time.
SOME VOLCANIC MATERIALS Magmas vary in their dissolved-gas (volatile) content and viscosity. Low-water-content, low-viscosity magma that reaches the surface typically moves as lava flows, with easy gas escape yielding nonexplosive eruptions. High- water- content, high-viscosity magma holds its volatiles, making gas escape difficult. Gas is forced to burst out of the magma, yielding explosive eruptions. Gas blasting into the atmo- sphere takes along chunks of magma and older rock known as pyroclastic debris ( pyro � fire; clastic � fragments).
Nonexplosive Eruptions Lava flows are especially typical of basaltic magma and exhibit a variety of textures ( table 6.6 ). Highly liquid lava may cool with a smooth, ropy surface called pahoehoe, (pronounced pa-Hoy-Hoy) ( figure 6.17 ). Slower-flowing, more viscous lava commonly has a rough, blocky texture called aa (pronounced ah-ah) ( figure 6.18 ).
Figure 6.17 Small-scale pahoehoe near Halemaumau, Hawaii. Photo by Pat Abbott.
Figure 6.18 Aa flow in Hawaii. Photo by Pat Abbott.
TABLE 6.6 Volcanic Materials
Lava Pahoehoe Smooth, ropy surface
Aa Rough, blocky surface
Pillow Ellipsoidal masses formed in water
Pyroclastic Air-fall Fragments
Fine ash (dust) Flour-size material
Coarse ash Sand size
Cinders Marble to baseball size
Blocks Big fragments, solid while airborne
Bombs Big fragments, liquid while airborne
Volcanic tuff Rock made of smaller fragments (e.g., deposit of a hot, gas-charged flow)
Volcanic breccia Rock made of coarse, angular fragments (e.g., deposit of a water-charged debris flow)
Glass Obsidian Nonporous glass
Pumice Porous glass (froth)
Explosive Eruptions Gaseous explosions break rock and tear apart magma and older rock into pyroclastic debris with a wide range of sizes, from dust to huge blocks and bombs ( table 6.6 ; figure 6.19 ). Airborne pyroclasts have their coarsest grains fall from the atmosphere first, closest to the volcano, followed by progres- sively finer material at greater distances away ( figure 6.20 ). An air-fall deposit can be recognized by the sorting of pyroclasts into layers of different sizes. Pyroclastic debris also can be blasted out over the ground surface as high-speed, gas-charged flows that dump material quickly, producing indistinct layer- ing and little or no sorting of the various-size particles.
Magma reaching the surface can solidify so quickly that crystallization cannot take place because there is no time for atoms to arrange themselves into the ordered atomic struc- tures of minerals. When magma cools this fast, it produces glass ( table 6.6 ). Cooled volcanic glass is known as obsidian ( figure 6.21 ). When gas escapes quickly and violently from lava, it may produce a frothy glass full of holes left by former gas bubbles; this porous material, known as pumice, contains so many holes it can float on water ( figure 6.21 b ). Scoria are rough crusts or chunks of basaltic rock full of holes made by expanding gases before solidification ( figure 6.21 c ).
abb22878_ch06_143-173.indd 154abb22878_ch06_143-173.indd 154 14/12/12 1:57 PM14/12/12 1:57 PM
G A R R E T T , M E G A N 1 3 2 4 T S
How a Volcano Erupts 155
(a)
(b)
(c)
Figure 6.21 Volcanic glass and scoria. (a) Obsidian is a dense, dark glass. (b) Pumice, a porous, light glass, floats in water. (c) Scoria, a basaltic rock with large pores, sinks in water. Photos (a) and (b) from NASA; (c) by Pat Abbott.
(a)
(b)
Figure 6.19 (a) This large blob of magma cooled while airborne and fell as a volcanic bomb , Irazu Volcano, Costa Rica. (b) Pyroclastic bombs kill people every year. (a) Photo by Pat Abbott. (b) Drawing by Jacobe Washburn.
Figure 6.20 Volcanic ash covers a house near Mount Pinatubo in the Philippines, June 1991. Photo by R. P. Hoblitt, US Geological Survey .
abb22878_ch06_143-173.indd 155abb22878_ch06_143-173.indd 155 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
156 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Side Note How a Geyser Erupts The eruption of water superheated by magma is called a geyser. The name is from the Icelandic word geysir, meaning “to gush or rage.” Geyser areas include Iceland, Chile, Yellowstone Park in the United States, North Island of New Zealand, and the Kamchatka Peninsula of Russia. All of these sites share common characteristics: subsurface water is present, and heat is abundant. Water from snow, rain, streams, and lakes is pulled below the ground surface by gravity, where it slowly moves through the network of voids presented by fractures and cavities, or pores, in rocks. The down- ward-circulating water encounters heat from a body of magma, absorbs heat, and then erupts ( figure 6.22 ).
This simple description ignores the complex interplay of tem- perature and pressure, which combine to set off an eruption. Water boils at 100°C (212°F) at sea level. At the 2,150 m (more than 7,000 ft) elevation of Yellowstone Park, with the reduced pressure of a thinner atmosphere above it, water boils at 93°C (193°F). However, water encountered while drilling 332 m (1,088 ft) below the surface at Norris Geyser Basin in Yellowstone Park was still liq- uid at 241°C (465°F) because the pressure at that depth is too great for it to change to steam.
When superheated water at depth does boil, its volume expands as liquid changes to steam, helping lift surrounding water upward to lower pressure levels, where more water flashes to steam helping lift more superheated water to lower pressure levels, and so on ( figure 6.23 ).
Reduction of pressure on superheated water causes it to change from liquid to gas, triggering the geyser eruption, analo- gous to the reduction of pressure that causes hot rock to change from solid to liquid, triggering a volcano eruption. A geyser erup- tion usually follows this sequence of events: (1) At depth, super- heated water flows out of tiny pressurized cracks into geyser reservoirs of larger volume; (2) as water temperature rises, some water flashes to steam; (3) the steam bubbles rise to lower pressure
Figure 6.22 Eruption of a geyser at Rotorua, North Island, New Zealand. Photo by Pat Abbott.
Elevation in feet
8,500
5,000
–5,000
–10,000
–15,000
Sea level
65˚C (150˚F)
He at
Heat
Heat
1,500–2,000˚F Magma
260˚C (500˚F)
93˚C (200˚F)
Figure 6.23 Surface water is pulled below ground by gravity, flows through holes in rocks, is superheated by magma, and through complex reductions in pressure, erupts at the surface as geysers.
levels, expanding continuously; (4) steam and bubbles become so abundant that they overwhelm the water, carrying it upward to levels of lower pressure, causing continual conversion to steam along the upward route; and (5) finally—the spectacular eruption.
abb22878_ch06_143-173.indd 156abb22878_ch06_143-173.indd 156 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 157
The Three Vs of Volcanology: Viscosity, Volatiles, Volume We can understand volcanoes anywhere in the world using the three Vs of volcanology: viscosity, volatiles, volume. Viscosity may be low, medium, or high, and it controls whether magma flows away or piles up. Volatile abundance may be low, medium, or high, and volatiles may ooze out harmlessly or blast out explosively. Volume of magma may be small, large, or very large. Volume correlates fairly well with eruption intensity; the greater the volume, the more intense the eruption.
Consider the five eruptive styles in figure 6.16 in terms of viscosity and volatiles ( table 6.7 ). The lower the volatile content and the viscosity, the more peaceful the eruption. As the volatile content increases, so can the explosiveness of the eruptions.
Applying what we have learned about magmas allows us to see linkages between eruptive behaviors and the land- forms built by volcanic activity. By mixing and matching the values among the three Vs, you can define volcanic land- forms ( table 6.8 ) and forecast the eruptive styles that occur at each of them.
TABLE 6.7 Eruption Styles and Explosiveness
Eruption Style Viscosity Volatiles
Volcanic Explosivity Index
Icelandic Low Low 0–1 (very low)
Hawaiian Low Low 0–1 (very low)
Strombolian Moderate Moderate 1–3 (low)
Vulcanian High High 2–5 (high)
Plinian High High 3–8 (very high)
TABLE 6.8 Volcanism Control by the Three Vs (Viscosity, Volatiles, Volume)
Viscosity � Volatiles � Volume � Volcanic Landforms Low Low Large Shield volcanoes
Low Low Very large Flood basalts
Low/medium Medium/high Small Scoria cones
Medium/high Medium/high Large Stratovolcanoes
High Low Small Lava domes
High High Very large Calderas
SHIELD VOLCANOES: LOW VISCOSITY, LOW VOLATILES, LARGE VOLUME The rocks of shield volcanoes form mostly from the solidi- fication of lava flows of basalt. These lava flows are low viscosity, contain less than one weight % volatiles, and are so fluid that they travel for great distances, somewhat analo- gous to pouring pancake batter on a griddle. Each basaltic flow cools to form a gently dipping, relatively thin volcanic rock layer. Many thousands of these lava flows must cool on top of each other over a long time to build a big volcano. A shield volcano, such as Mauna Loa in Hawaii, has a great width compared to its height, whereas a volcano built of high- viscosity magma, such as Mount Rainier in Washing- ton, has a great height compared to its width ( figure 6.24 ).
Hawaiian-type Eruptions As with virtually all volcanic eruptions, Hawaiian-type eruptions are commonly preceded by a series of earthquakes as rock fractures and moves out of the way of swelling magma. When these fractures split the ground surface, they suddenly reduce pressure, allowing gas to escape from the top of the magma body. This can create a beautiful “curtain of fire” where escaping jets of gases form lines of lava foun- tains up to 300 m (1,000 ft) high. Also common in the Hawaiian eruption is formation of a low cone with high fountains of magma. After the initial venting of gas, great floods of basaltic lava spill out of the fissures and flow down the mountain slopes as red-hot rivers ( figure 6.25 ). These eruptions may last from a few days to a year or more. Although few lives are lost to Hawaiian volcanism, the ubiq- uitous lava flows engulf and incinerate buildings, bury high- ways, cause drops in property value of homes near the latest flow, and cause some homeowners to lose their peace of mind ( figure 6.26 ).
Hawaiian volcanoes capable of eruption include Hale- akala on the island of Maui, the five volcanoes that make up the island of Hawaii, and the growing but still subsea volcano of Loihi. Haleakala last erupted around 1790. Today, its 49 km 2 (19 mi 2 ) summit caldera is a major tourist attraction.
abb22878_ch06_143-173.indd 157abb22878_ch06_143-173.indd 157 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
158 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
HAWAIIAN ISLANDS
Kohala
Hualalai
Mauna Loa Kila uea
Mauna Kea
1859
1801
180 1
1750
1855
1969-74 1983 and
continuing
1840
1793 1977
1750
1955
1960
1949
1950
1926
191 6
19 07
19 07
18 87
18 68
1949
1942
1852
188 0
1961 -69
1832 1823
1950
1942
19 35 1
84 3
18 99
194 0
19 20
Kauai
Niihau Oahu
Lanai
Molokai
Maui
Hawaii
Loihi
0 10 20 mi
Figure 6.25 Map of Hawaii showing some historic lava flows. The circular color overlay shows boundaries of mantle plume rising from the hot spot at depth.
Figure 6.26 Lava flows caused the Wahalua Visitor’s Center in Hawaii Volcanoes National Park to burn to the ground in 1989. Photo by J. D. Griggs, US Geological Survey .
Mauna Loa Kilauea Ocean5.8 km (19,000 ft) deep
Mount Rainier
190 km (120 miles)
Figure 6.24 A shield volcano, such as Mauna Loa in Hawaii, has a great width compared to its height. A stratovolcano, such as Mount Rainier in Washington, has a great height compared to its width. Data source: Tilling, R. I., et al., Eruptions of Hawaiian Volcanoes, US Geological Survey , 1987.
In the last 200 years, eruptions have occurred only on the three southernmost volcanoes on the island of Hawaii and below sea level on Loihi. The island-to-be (Loihi) is located about 30 km (19 mi) off the southeastern shore of Hawaii. Loihi’s peak is about 969 m (3,175 ft) below sea level, and the weight of the overlying ocean water suppresses the explosiveness of the eruptions for now, but the volcano is building upward impressively.
In general, the volcanism on Hawaii is relatively peace- ful and acts as a magnet attracting tourists to witness nature’s spectacle. But there are exceptions to this statement.
Killer Event of 1790 Although less than 0.5% of Hawaiian magma is blown out as pyroclastic material, rare killer events do occur. In 1790, trav- eling parties from King Keoua’s army were caught and many of the people killed by a blast from Kilauea Volcano. The army was passing through the area but was stopped by eruptions. After three days of waiting, it split into three parties of about 80 people each. As the parties marched southwest down the trail from Kilauea, disaster struck. An explosion column burst upward, with a base surge sweeping outward as a dense, basal cloud. Base surges can travel at hurricane speeds as masses of ground-hugging hot water and gases with or without magma fragments. The base surge in 1790 overtook King Keoua’s middle party, killing them all. The victims huddled together, grasping each other to withstand the hurricane-force blast, but the hot gases seared their lungs and the intense heat scorched their skin. The base surge caught up with the lead party, but it had weakened, allowing most of those people to survive. The trailing party was alongside the blast and suffered no deaths or
abb22878_ch06_143-173.indd 158abb22878_ch06_143-173.indd 158 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 159
In Greater Depth Volcanic Explosivity Index (VEI) How often do big volcanic eruptions occur? On average, about once every three years, according to the volcanic explosivity index (VEI). Combining the historic record with the geologic information stored in the rock record, the major volcanic eruptions occurring between the years 1500 and 1980 were evaluated for their size. Factors evaluated include (1) volume of material erupted, (2) how high the eruption column reached, and (3) how long the major eruptive burst lasted ( table 6.9 ). During the 481-year interval stud- ied, 126 major eruptions occurred, with the number increasing in
modern times. The increase in big eruptions in the 19th and 20th centuries is certainly due to better reporting of events, rather than an actual increase in major eruptions.
The VEI ranges from 0 to 8. The biggest event since 1500 CE was the VEI 7 eruption of Tambora in 1815 in Indonesia. This erup- tion caused a cooling of the world climate during the following year (see chapter 12). Four VEI 6 events occurred in the 481-year period, including the 1883 eruption of Krakatau, also in Indonesia (described in the section on calderas in this chapter). Volcanic events with high VEI values are those of Vulcanian and Plinian-type eruptions. A fifth VEI 6 event occurred in 1991 when Mount Pina- tubo erupted.
TABLE 6.9 Volcanic Explosivity Index (VEI)
VEI
0 1 2 3 4 5 6 7 8 Volume of ejecta (m 3 ) <10 4 10 4 −10 6 10 6 −10 7 10 7 −10 8 10 8 −10 9 10 9 −10 10 10 10 −10 11 10 11 −10 12 >10 12
Eruption column height (km) <0.1 <0.1−1 1−5 3−15 10−25 >25
Eruptive style <-----Hawaiian-----> <-----Vulcanian----->
<-----Strombolian-----> <------------Plinian------------>
Duration of continuous blast (hours)
<-----<1-----> <-----1–6-----> <-------->12-------->
<-----6–12----->
Source: After Newhall and Self (1982).
injuries. Although basaltic magma is not likely to explode, this case history shows how it can incorporate water into the magma and heat groundwater to cause an eruption, including a base surge of superheated steam. The 1790 event is worth remembering for today’s watchers of Hawaiian eruptions—at least seek the high ground during your viewing.
FLOOD BASALTS: LOW VISCOSITY, LOW VOLATILES, VERY LARGE VOLUME Flood basalts are the largest volcanic events known on Earth. Two important characteristics are (1) the immense amounts of mass and energy they pour onto Earth’s surface, and (2) their geologically brief duration. Flood basalts erupt tremendous volumes of magma within a geologically short time—for example, 1 to 3 million years. Hot spots also bring up huge volumes of magma but do so during a long period—for exam- ple, 100 million years.
Eruptions from individual volcanoes transfer a lot of heat from Earth’s interior to the surface, but the most impressive movements of heat occur with flood basalts. The numbers that describe the volumes of magma they erupt and the
surface areas they bury with lava are so large they are hard to visualize. For example, 252 million years ago, up to 3 million km 3 (800,000 mi 3 ) of basalt flowed out and covered almost 4 million km 2 (1.5 million mi 2 ) of Siberia, Russia. Visualize basalt flows covering an area measuring about 1,200 mi by 1,200 mi with lava tens of meters thick. How does this area compare with the area of your state or province? Visualize your entire state or province buried beneath lava tens of meters thick.
Flood basalts occur on all continents and on all ocean floors, but none has occurred in historic times. Flood basalt eruptions obviously devastate a region, but can they have global effects? Yes, not from the lavas directly, but from the climate-modifying volatiles such as carbon dioxide (CO 2 ) or sulfur dioxide (SO 2 ) they release into the atmosphere. Is it a coincidence that the greatest mass extinction known occurred during the time the Siberian flood basalt was being erupted? Probably not.
Another flood-basalt episode occurred 65 million years ago, when about 1.5 million km 3 (360,000 mi 3 ) of basalt flowed out and covered about 1.5 million km 2 (580,000 mi 2 ) of the Deccan region of India. This also coincides with a mass extinction, the famous one that includes non-avian dinosaurs.
abb22878_ch06_143-173.indd 159abb22878_ch06_143-173.indd 159 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
160 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
SCORIA CONES: MEDIUM VISCOSITY, MEDIUM VOLATILES, SMALL VOLUME Scoria cones are conical hills, typically of low height, formed of basaltic to andesitic pyroclastic debris piled up next to a volcanic vent. Scoria cones commonly are produced during a single eruptive interval lasting from a few hours to several years. The scoria, or cinder cone, has a summit crater, the basin on top of the cone that is usually less than 2 km (1.2 mi) in diameter. The summit crater may hold a lava lake during eruption. After the excess gas has been expelled from the magma body, the lava may drain and emerge from near the base of the cone. When that eruption ceases, scoria cones usually do not erupt again.
Strombolian-type Eruptions Scoria cones are built mainly by Strombolian-type eruptions (see figure 6.16 ). The volcano Stromboli, offshore from southwestern Italy, has had almost daily eruptions for millen- nia. Its central lava lake is topped by a cooled crust. Even the tidal cycle disrupts the lava-lake crust, thus triggering erup- tions. Gas pressure builds quickly beneath the crust, and eruptions occur as distinct and separate bursts up to a few times per hour. Each eruption tosses pyroclasts tens to hun- dreds of meters into the air. For many centuries, tourists have climbed Stromboli to thrill at the explosive blasts, but usually every year, a few of those tourists die when hit by large pyro- clastic bombs. Strombolian eruptions are not strong enough to break the volcanic cone.
On 20 February 1943, a new volcano was born as erup- tions blasted up through a farm field near the village of Paricutin in the state of Michoacan, Mexico ( figure 6.27 ). The volcano erupted for nine years, building a distinctive scoria cone. Pyroclastic debris and lava flows buried about 260 km 2 (100 mi 2 ) of land and destroyed the towns of Paricutin and San Juan de Parangaricutiro.
STRATOVOLCANOES: HIGH VISCOSITY, HIGH VOLATILES, LARGE VOLUME Stratovolcanoes, or composite volcanoes, commonly are steep-sided, symmetrical volcanic peaks built of alternating layers of pyroclastic debris successively capped by high- viscosity andesitic to rhyolitic lava flows that solidify to form protective caps. Stratovolcanoes may show marked varia- tions in their magma compositions from eruption to eruption, and their eruptive styles include Vulcanian and Plinian. Some of Earth’s most beautiful mountains are stratovolcanoes—for example, Mount Kilimanjaro in Tanzania, Mount Shasta in California, Mount Rainier in Washington, and Mount Fuji in Japan ( figure 6.28 ).
Vulcanian-type Eruptions All volcanoes take their name from Vulcan, the Roman god of fire and blacksmith for the gods. The prototypical volcano
Figure 6.27 Paricutin Volcano erupting in 1943. This 400 m (1,300 ft) high scoria cone is in the state of Michoacan, Mexico. Photo from US Geological Survey .
is one of the Aeolian Islands in the Tyrrhenian Sea north of Sicily. The fire and smoke emitted from the top of the moun- tain reminded observers of the chimney of Vulcan’s forge, so the mountain was named Vulcano. Vulcanian eruptions alter- nate between thick, highly viscous lavas and masses of pyro- clastic material blown out of the volcano. Some Vulcanian eruptions are more violent blasts of high-viscosity magma loaded with trapped gases. The material blown out during eruptions covers wide areas. Vulcanian eruptions commonly are the early phase in the eruptions of other volcanoes as they “clear their throats” before emitting larger eruptions.
Plinian-type Eruptions Plinian eruptions are named after the 17-year-old Pliny the Younger in honor of his detailed written observations of the 79 ce eruptions of Vesuvius that claimed the life of his well-known uncle Pliny the Elder. In Plinian eruptions, the volcano “throat is now clear,” and incredible gas-powered vertical eruption columns carry pyroclastic debris, including
abb22878_ch06_143-173.indd 160abb22878_ch06_143-173.indd 160 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 161
Figure 6.28 Mount Fuji, a symmetrical stratovolcano rising 3,776 m (12,385 ft) above sea level, Honshu, Japan. The last major eruptions were in 1707–1708. Photo © Corbis RF.
Bay of Naples
Nisida
Procida
Ischia
Cuma
Pozzuoli Baia
Misenum Cape Miseno
Capo di Monte
Herculaneum
Torre del Greco
Torre del Annunziata
Castellammare StabiaeStabiae
Pompeii
200 cm
100 cm
50 cm
Somma
Nola
San Sebastiano Portici Resina
Sor rent
o Pe nins
ula
Sor rent
o Pe nins
ula
CapriCapri
Campi PhlegraeiCampi Phlegraei
Naples
10 cm Ottaiano
Mount Vesuvius
0 5 10 mi
Figure 6.29 Map of the Bay of Naples area showing the location of Mount Vesuvius. Pumice fallout from the 79 CE eruption is contoured in centimeters. Pompeii and Stabiae were buried by pumice; Herculaneum by lahars.
lots of pumice, up to 50 km (30 mi) into the atmosphere (see figure 6.16 ). The Plinian eruption is a common final phase in a major eruptive sequence. About two to three Plinian erup- tions occur each century.
Vesuvius, 79 ce Vesuvius began as a submarine volcano in the Bay of Naples. It grew greatly in size, and its rocky debris filled in the waters that once separated it from mainland Italy ( figure 6.29 ). What is the cause of the volcanism at Vesuvius and the neigh- boring volcanoes of Stromboli, Vulcano, Etna, and others? The subduction of Mediterranean seafloor beneath Europe to make room for the northward charge of Africa.
In 79 ce , when Mount Vesuvius began erupting, most residents fled from Pompeii. Those who stayed first experi- enced volcanic ash clouds dropping pumice. Pompeii lay downwind and was buried by pumice fragments accumulat- ing up to 3 m (10 ft) deep ( figure 6.29 ). Researchers estimate that about 60% of the people who remained in Pompeii sur- vived the first flows of ash and pumice. About half of the survivors then fled, but many of them died when they were caught outside during later flows of ash and pumice. The people still inside houses remained alive, only to suffocate from breathing hot particles and gases seeping out of the vol- canic debris. Death was not always quick. Some bodies were found inside houses on top of thick layers of pumice, giving
abb22878_ch06_143-173.indd 161abb22878_ch06_143-173.indd 161 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
162 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
evidence of hours of struggle by people fighting to stay alive. Their hands held cloths over their mouths as they tried to avoid asphyxiation from gases seeping out of the pumice.
Many other people were found near the sea (see figure 6.6). They escaped the falling pumice, but ground- hugging pyroclastic flows, full of hot gases, finished them off (for a modern example of a major Vulcanian- type eruption, see figure 6.3 ). About 4,000 people died. The more-distant town of Stabiae was also mostly destroyed. It was here that Pliny the Elder died; the weak heart of the overweight man failed at age 56 under the stress of the farthest-reaching gas-rich flow.
Testing of rocks formed during the 79 ce eruption, as well as roof tiles from Pompeii, indicates that the cloud of volcanic ash and pumice that smothered Pompeii erupted out of Vesuvius at about 850°C (1,550°F) and then cooled to less than 380°C (710°F) by the time it reached the city. Roof tiles in Pompeii were heated to maximum tempera- tures of 340°C (640°F), while some walls on the partially protected down-flow side of houses reached temperatures of around 180°C (350°F), presumably because cooler air mixed into the volcanic ash cloud.
Following the Vulcanian-type eruption, the volcano entered a second phase, the Plinian phase, where it blew immense volumes of pyroclasts up to 32 km (20 mi) high in the atmosphere. The height of the eruption column varied as the volcanic energy waxed and waned. During weakened intervals, the great vertical column of ashes would temporarily collapse, sending surges and pyroclas- tic flows down the volcano slopes. Pompeii was buried under an additional 6 to 7 ft of pyroclastic debris. These were the flows that finished off the surviving Pompeiians.
A Plinian eruption not only blows ashes to great heights but also volcanic gases. Water, as abundant steam, can be blown high into the atmosphere, cooling and condens- ing and then falling back down as rain—heavy rain. Some volcanic eruptions create their own “weather.” Rain falling on thick piles of pyroclastic debris, sitting unstably on the steep slopes of Vesuvius, set off thick volcanic mudflows ( figure 6.30 ). Any gravity-pulled mass movements of muddy volcanic debris are known as lahars, an Indonesian word (for a modern example, see figure 6.31 ). Lahars buried the city of Herculaneum up to 20 m (65 ft) deep in pumice, ashes, and volcanic rock fragments jumbled together in a confused mass. However, this was during the second phase of the erup- tion, and most people had used the day or two before to clear out of the area, so the loss of life was not nearly as great as at Pompeii. Today, the town of Ercolano lies on top of the mud- flows burying Herculaneum. The lessons of history have not been well learned here.
Recent seismic surveys have helped define the magma body underlying Vesuvius today. Seismic waves with low- ered velocity define a 400 km 2 (150 mi 2 ) horizontal, broad sheet of partially molten rock. This magma body lies at a depth of 8 km (5 mi) below the surface. Assuming a magma
Condensed steam
Steam Ashes
Condensed steam
Vesuvius
Herculaneum
As h m
as s
Steam
Volcanic mudflow
Rain of ash
Figure 6.30 “Volcano weather” and the formation of lahars. Prolonged vertical eruption leads to accumulation of debris (ash mass) on steep slopes of the volcano. Steam blown upward into cold, high altitudes condenses and falls back as rain. The stage is set: steep slopes � loose volcanic debris � heavy rain � lahars. Some volcanic eruptions even generate their own lightning.
body thickness of 0.5 to 2 km (0.3 to 1.2 mi), the volume of magma is about 200 to 800 km 3 (50 to 200 mi 3 ). This magma reservoir is fed from below, and it can supply magma to smaller layers closer to the surface, as it has in fueling past eruptions. The millions of people around the Bay of Naples live in real danger (see figure 6.7 ).
LAVA DOMES: HIGH VISCOSITY, LOW VOLATILES, SMALL VOLUME Lava domes form when high-viscosity magma with a low content of volatiles cools quickly, producing a hardened dome or plug a few meters to several kilometers wide and a few meters to 1 km high. Lava domes can form in as little as a few hours, or they may continue to grow for decades. The formation of lava domes can be visualized as part of a larger eruptive process. When a large volume of hot rock/magma rises and undergoes decompression melting, dissolved vola- tiles are freed. Many of the freed gases rise and accumulate at or near the top of the magma mass. When a major eruption occurs, these gases power the initial Vulcanian-type blast and then the succeeding Plinian-type eruption, which lasts until the excess volatiles have escaped. What type of magma
abb22878_ch06_143-173.indd 162abb22878_ch06_143-173.indd 162 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 163
Figure 6.31 Lahars from a Vulcanian-type eruption of Mount Pinatubo destroyed bridges to Angeles City, Philippines, 12 August 1991. Photo by T. Casadevall, US Geological Survey.
Figure 6.32 The Novarupta lava dome formed as hardened magma plugged the central magma pipe of the 1912 eruption of Katmai Volcano in southern Alaska. The dome is 244 m (800 ft) across and 61 m (200 ft) high. Photo from US Geological Survey.
remains? Often it is a low-volatile, high-viscosity paste that oozes upward slowly and cools quickly, forming a plug in the throat of the volcano. Figure 6.32 shows the lava dome emplaced in Mount Katmai in southern Alaska following its 1912 eruption, the biggest eruption of the 20th century.
Lava domes can provide spectacular sights. After the 1902 eruptions of Mont Pelée in the Caribbean killed more than 30,000 people (see chapter 7), a lava dome formed as a great spine that grew about 10 m/day (33 ft/day) and rose above the top of the volcano. The spine of hardened magma was forced upward by the pressure of magma below until it stood more than 300 m (more than 1,000 ft) higher than the mountaintop, like a giant cork rising out of a bottle.
Do lava domes present hazards? Yes, in the 1990s, they were responsible for 129 deaths: 19 from Soufriere Hills Volcano on Montserrat in 1997, 66 from Mount Merapi in Indonesia in 1994, and 44 from Mount Unzen in Japan in 1991. The hardened, brittle lava dome rock can fail in a gravity- pulled landslide from the mountain, or magma trapped below the brittle lava dome can break out in a violent eruption.
A Typical Eruption Sequence A common pattern for a major eruptive episode is that gas-rich materials shoot out first as a Vulcanian blast, quickly followed by a longer-lasting, gas-driven Plinian eruption. When the gas is depleted, then gas-poor, high-viscosity magma slowly oozes out to build a lava dome over an extended period of time.
The volcanic sequence could be described as a Vulcanian precursor, a Plinian main event, and a lava dome conclusion.
CALDERAS: HIGH VISCOSITY, HIGH VOLATILES, VERY LARGE VOLUME Caldera-forming eruptions are the largest of the violent, explosive volcanic behaviors. Calderas are large volcanic depressions formed by roof collapse into partially emptied magma reservoirs. Calderas differ from volcanic craters. Both are topographic depressions, but a crater is less than 2 km (1.2 mi) in diameter and forms by outward explosion. Calderas are larger; they range from 2 to 75 km (45 mi) in diameter and form by inward collapse ( figure 6.33 ).
Calderas form in different settings: (1) Calderas that form at the summits of shield volcanoes include those at Mauna Loa and Kilauea on the island of Hawaii. A recent subsidence of the Kilauea caldera followed draining of magma to feed eruptions out of lower-elevation rift zones. (2) Calderas forming at the summits of stratovolcanoes include Crater Lake in Oregon, Krakatau in Indonesia, and Santorini in the Aegean Sea ( figure 6.34 a,b,c ). Caldera collapses occurred following sustained Plinian eruptions of 55 km 3 (13 mi 3 ) of pyroclasts at Crater Lake, 10 km 3
abb22878_ch06_143-173.indd 163abb22878_ch06_143-173.indd 163 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
164 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Side Note British Airways Flight 9 On 24 June 1982, 247 people on British Airways flight 9 boarded a Boeing 747 for a night flight from Kuala Lumpur, Malaysia, to Perth, Australia. The night was moonless but clear, and the weather forecast was good. The crew took the big airplane up to its cruising altitude of 37,000 ft and then relaxed a bit. Weather radar showed that out- side conditions were normal. But the pilot noticed puffs of “smoke” and an acrid or electrical odor. As he sat in the pilot’s seat and peered through the front windscreens, the atmosphere seemed to be on fire as intense electricity danced about. Out the side windows, the engines were glowing as if they were lit inside. Then the flight engi- neer called out: “Engine failure number 4,” followed shortly by:
“Engine failure number 2. “Three’s gone. “They’ve all gone.” The pilot thought, “Four engines do not fail.” The instrument
panel was contradictory; some gauges read normal while others told of problems with a confusing lack of pattern.
Meanwhile, the plane was descending slowly. At 26,000 ft, the oxygen masks were released, but some didn’t work; in this emer- gency, a steep descent was initiated to get down to atmospheric levels with more oxygen. When the plane reached 14,000 ft, the pilot said:
“Good evening ladies and gentlemen. This is your captain speaking. We have a small problem. All four engines have stopped. We are doing our darndest to get them going again. I trust you are not in too much distress.” His words could not have brought much comfort to those passengers in window seats who had been watch- ing the engines that seemed to be on fire.
What to do? Land on the ocean during a dark night? Too dangerous. Finally, at 12,000 ft, engine number 4 started, and 90 seconds later, the other three engines started. The pilot set the plane to climbing to avoid hitting the mountainous Indonesian topography, but at 15,000 ft, the bad atmospheric problems began again. Descent was once again initiated. Permission was granted for an emergency landing at Jakarta, but the approach was hazardous. The front and side windows were frosted and opaque, so the co-pilot had to look out a little side window and give instructions to the pilot landing the huge, fast-moving plane. At last, the plane landed smoothly, and the passengers cheered and clapped.
What happened that night to BA9? It flew during an eruption of Mount Galunggung and passed through its seething cloud of hot volcanic ash and larger pyroclastic debris. The volcanic ash clogged the engines, frosted the windscreen, and turned BA9 into a terror-filled flight. Airplanes must avoid volcanoes in eruption.
Figure 6.33 A crater (less than 2 km across) formed atop the volcano in the foreground during eruptions. The large caldera (more than 2 km across) at low elevation in the background formed when its volcano collapsed during a massive eruption, Kamchatka Peninsula, Russia. Photo from US Geological Survey.
(2.4 mi 3 ) at Krakatau, and 40 km 3 (10 mi 3 ) at Santorini. These eruptions opened void spaces that caused mountain peaks to collapse into their magma chambers. (3) Giant con- tinental calderas are huge negative landforms such as Lake Yellowstone in Wyoming or Long Valley in California. These broad and deep depressions formed following the rapid
eruption of 2,000 km 3 (475 mi 3 ) of pyroclasts at Yellowstone and 600 km 3 (140 mi 3 ) at Long Valley. The huge volumes of magma pour out in short amounts of time as ultra-Plinian eruptions with extra-high ash columns and widespread sheets of outward-flowing ash and pumice.
The most recent example of an ultra-Plinian eruption occurred 74,000 years ago at Toba, on the island of Sumatra in Indonesia. The caldera at Toba is 30 km (20 mi) by 100 km (60 mi) long and has a central raised area inside it that is more than 1 km high. The raised area formed during the millennia following the giant eruption; this resurgent topog- raphy inside the caldera gives these features their name— resurgent calderas.
Crater Lake (Mount Mazama), Oregon Crater Lake is one of the jewels in the U.S. national park sys- tem. Its intense blue waters are pure and lie cradled in a high- rimmed, nearly circular basin. Crater Lake is about 9.5 km (6 mi) across and as deep as 589 m (1,932 ft) ( figure 6.35 ).
Several thousand years ago, the stratovolcano Mount Mazama stood about 3,660 m (12,000 ft) high as one of the Cascade Range volcanoes above the Cascadia subduction zone ( figure 6.36 a ). More than 7,600 years ago, a major eruption began blowing sticky magma out of the mountain as glassy, gas-bubble-filled pumice and ashes ( figure 6.36 b ). The magma had too high a viscosity to flow as a liquid, so it erupted as pyroclastic flows and Plinian columns. As the erupted mate- rial grew in volume, its debris covered much of the U.S. Pacific Northwest and part of Canada with a thick, distinctive ash layer that is easily recognizable. Mazama ash is found
abb22878_ch06_143-173.indd 164abb22878_ch06_143-173.indd 164 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 165
Pumice Desert
–589 m
Wizard Island
Phantom ship
Visitor center
(a) Crater Lake, Oregon
0 5 10 km
0 3 6 mi
1883 collapse Lang
Verlaten Anak
Krakatau
Rakata
(b) Krakatau, west of Java
Thera
–390 m
Kameni
T he
ra si
a
Aspronisi
1628 BCE collapse
Akrotiri
(c) Santorini, Aegean Sea
Figure 6.34 Map of some collapse calderas. (a) The nearly circular caldera of Crater Lake, Oregon, formed about 5677 BCE . (b) Island remnants of old volcano Krakatau; crudely ovoid shape of the 1883 collapse; and the new and growing volcano, Anak Krakatau. (c) Caldera in volcano Santorini that collapsed into the Aegean Sea about 1628 BCE .
Figure 6.35 Crater Lake, Oregon, fills the caldera of Mount Mazama, which collapsed in the year 5677 BCE . Wizard Island is visible. Photo © Robert Glusic/Getty Images RF.
in the Greenland glacier within the ice layer formed during the snowfall season of the year 5677 bce . About 40 km 3 (10 mi 3 ) of magma was ejected. Evacuation of this immense volume of magma left so tremendous a void below the sur- face that the weakened mountain peak collapsed and moved down in pistonlike fashion into the emptied magma chamber ( figure 6.36 c ). The collapse produced a caldera about 10 km (6 mi) across that has collected the water for Crater Lake and hosted the growth of a 1,000-year-old successor volcanic cone called Wizard Island ( figures 6.34 a and 6.35 ).
The eruption of Mount Mazama affected American Indians, as evidenced by moccasin tracks and artifacts found beneath the distinctive ash layer. What have caldera-forming collapse events wrought elsewhere?
Krakatau, Indonesia, 1883 Today, Krakatau (Krakatoa) is a group of Indonesian islands in the Sunda Strait between Sumatra and Java (see lower right
abb22878_ch06_143-173.indd 165abb22878_ch06_143-173.indd 165 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
166 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
figure 4.10). It is part of the grand arc of volcanoes built above the subducting Australia-India plate. Krakatau is a big strato- volcano that builds up out of the ocean and then collapses. Its larger outline is still distinguishable (see figure 6.34 b ).
From the ruins of an earlier collapse, magmatic activity built Krakatau upward through the 17th century. After two centuries of quiescence, volcanic activity resumed on 20 May 1883. By August 1883, moderate-size Vulcanian eruptions were occurring from about a dozen vents. At 2 p.m. on 26 August, a large blast shot volcanic ashes and pumice 28 km (17 mi) high as one of the cones collapsed into the sea, setting off huge tsunami. Eruptions were so noisy that night that sleep was not possible in western Java, including the capital city of Djakarta (then called Batavia). The early morning hours of 27 August were rocked by more ear- hammering eruptions, and further volcanic collapses sent more giant tsunami to wrack the coastal villages. Day was turned to nightlike dark- ness as heavy clouds of volcanic ashes blocked the sunlight. At 10 a.m., a stupendous blast rocketed a glowing cloud of incan- descent pumice and ashes 80 km (50 mi) into the atmosphere. This blast was distinctly heard 5,000 km (3,000 mi) away.
The 10 a.m. volcano collapse sent tsunami higher than 35 m (115 ft) sweeping into bays along the low coastlines of Java and Sumatra. The volcanic eruptions caused tsunami that destroyed 295 towns and smashed or drowned an esti- mated 36,000 people. The eruption sequence blew out 18 km 3 (4 mi 3 ) of material (95% fresh magma and 5% pulverized older rock), creating a subterranean hole into which 23 km 2 (5.5 mi 2 ) of land collapsed. Where islands with elevations of 450 m (1,476 ft) had stood, there now was a hole in the sea- floor 275 m (900 ft) deep.
The amount of magma erupted at Krakatau in 1883 was less than half that of the Mount Mazama eruption. But Krakatau collapsed into the sea, sending off tsunami.
In 1927, Krakatau began rebuilding a new volcanic cone called Anak Krakatau—“child of Krakatau”; it is still grow- ing (see figure 6.34 b ). We will hear more from Krakatau.
Santorini, Greece As the Mediterranean oceanic plate subducts beneath Europe, it causes numerous volcanoes. One of the biggest is the stratovolcano Santorini in the Aegean Sea. Today, Thera is the largest island in a circular group marking the sunken remains of Santorini (see figure 6.34 c ). Thera is one of the most popular tourist sites in the Greek islands ( figure 6.37 ), but around 1628 bce , Santorini underwent an explosive series of eruptions that buried the Bronze Age city of Akrotiri on Thera to depths of 70 m (230 ft) in four distinct phases. Where there had been a large island made of several volcanic cones, there now exists a huge caldera with depths of 390 m (1,280 ft) below sea level.
Caldera-forming eruptions are low-frequency, high- impact events. Santorini had a major eruption more than 21,000 years ago, then about 18,000 years passed without a huge eruption—until the ultra-Plinian eruption and volcano collapse during the Minoan civilization. A recent study of
Mount Mazama 3,660 m (12,000 ft) elevation
Magma
(a)
(b)
(c)
(d)
Magma
Magma
Magma
Crater Lake
Collapse
Eruption in 5677 BCE
Figure 6.36 How Crater Lake formed. (a) The Mount Mazama volcano stood high. (b) A gaseous eruption in the year 5677 BCE emptied a huge volume of viscous magma. (c) The gigantic eruption left a void inside the weakened mountain, and the unsupported top fell down into the emptied magma chamber. (d) The waters of Crater Lake now fill the caldera, and a small new volcanic cone (Wizard Island) rises above lake level.
abb22878_ch06_143-173.indd 166abb22878_ch06_143-173.indd 166 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 167
Figure 6.37 Aesthetic view of the inside of the Santorini caldera. A classic white-washed Greek church and some homes have been built on the cliff edge. Photo © Adam Crowley/Getty Images RF.
chemically zoned crystals of plagioclase that grew in the 855 o C (1,570 o F) rhyolitic magma reservoir determined how much time it took to recharge the magma reservoir before the overwhelming eruption of 1628 bce. The results are surpris- ing. After 18,000 years of waiting, it took less than 100 years to supply the 40–60 km 3 (10–15 mi 3 ) of magma that were erupted. The huge supply of viscous magma forced relatively small-volume dikes or columns of magma to the surface, initiating the pressure drop that triggered the eruption.
It is sobering for us today to consider that a civilization- changing mega-eruption might take only a few decades of magma recharge before occurring.
Yellowstone National Park A giant continental caldera exists in Yellowstone above a hot spot, a long-lived mantle plume that the North American continent is drifting across. The hot spot occupies a relatively fixed position above which the North American plate moves southwestward about 2 to 4 cm/yr (0.8 to 1.6 in/yr). Plate movement over the hot spot during the last 15 million years is recorded by a trail of surface volcanism cut across the Snake River plain in Idaho and on into Wyoming ( figure 6.38 ). At present, Yellowstone National Park sits above the hot spot, and a large body of rhyolitic magma lies about 5 to 10 km (3 to 6 mi) beneath it.
In the past 2 million years, three catastrophic ultra- Plinian eruptions have occurred at Yellowstone at 2 million, 1.3 million, and 0.6 million years ago ( figure 6.39 ). Such mega-eruptions do not come often, but in a few short weeks, they pour forth virtually unimaginable volumes of rhyolitic magma, mostly as pyroclastic flows. The oldest event erupted 2,500 km 3 (600 mi 3 ) of magma, the middle one emptied 280 km 3 (70 mi 3 ), and the youngest dumped out 1,000 km 3 (240 mi 3 ). (Compare these magma volumes to the 1980 erup- tion of Mount St. Helens, which totaled 1 km 3 .) An eruption of 1,000 km 3 of rhyolitic pyroclastic flows would cover a surrounding area of 30,000 km 2 (11,500 mi 2 ) with a mass of pyroclastic debris ranging from a few to more than 100 m in thickness. The weight of volcanic material would cause a 500 km 2 (200 mi 2 ) area to sink isostatically.
The Yellowstone mega-eruption of 600,000 years ago created a giant caldera that is 75 km (47 mi) long and 45 km (28 mi) wide. Look again at figure 6.39 and consider the size of the giant caldera and the extent of its emitted pyroclastic flows: in a matter of days, all life in the area would have died and been deeply buried.
Eruptive Sequence of a Resurgent Caldera Giant caldera-forming eruptions go through a characteristic sequence. They begin when a very large volume of rhyolitic magma rises to within a few kilometers below the surface, bowing the ground upward ( figure 6.40 a ). The magma body accumulates a cap rich in volatiles and low-density compo- nents such as SiO 2 .
A mega-eruption begins with a spectacular circular ring of fire as Plinian columns jet up from circular to ovoid frac- tures surrounding the magma body ( figure 6.40 b ). The escaping magma erodes the fractures, thus increasing the size of the eruptive vents so that more and more magma escapes.
As greater volumes of gas “feel” the lessening pressure, the magma begins gushing out of the fractures in mind- boggling volumes ( figure 6.40 c ). The outrushing magma is too voluminous to all go airborne, so most of it just pours away from the vents as pyroclastic flows, the fastest way to remove gas-laden, sticky magma.
As the subsurface magma body shrinks, the land surface sinks as well, like a piston in a cylinder, creating a giant caldera ( figure 6.40 d). The removal of 1,000 km 3 (240 mi 3 ) of magma creates a void, an isostatic imbalance, that is filled by a new mass of rising magma that bows up the caldera floor to create a resurgent dome ( figure 6.41 ). Resurgent domes may be viewed as the reloading process whereby magma begins accumulating toward the critical volume that will trigger the next eruption.
The areas alongside resurgent domes are commonly occupied by lakes (see figure 6.39 ). Imagine driving the many miles from Yellowstone Lake to Old Faithful Geyser, all the time staying within the gigantic collapse caldera of the 600,000-year-old eruption. When will the next mega- eruption occur?
abb22878_ch06_143-173.indd 167abb22878_ch06_143-173.indd 167 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
168 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
15 30 mi0
50 km10 20 30 400
MONTANA
WYOMING
IDAHO
Yellowstone National Park
Resurgent dome
Shoshone Lake
Caldera Rim
Old Faithful
0.6 m .y.
2.0 m
.y.
Island ParkSnake River
Plain
Hebgen Lake
Heart LakeLewis
Lake
Yellowstone Lake
1. 3
m .y
.
Figure 6.39 The Yellowstone hot-spot area. The North American plate is moving southwest, and thus the hot-spot magma plume erupts progressively farther northeast with time. Three giant calderas have erupted in the past 2 million years—at 2, 1.3, and 0.6 million years ago. The wiggly-lined area was covered by hot, killing pyroclastic flows during the eruption of 600,000 years ago. Notice the resurgent domes.
Ashfall Nebraska
Yellowstone
10.3
12.5
6.3
Figure 6.38 Track of the Yellowstone hot spot in western North America. See the 600 km (370 mi) long trail marking the path of the North American plate. Today, the hot spot underlies Yellowstone National Park. Numbers shown are millions of years since each site was over the hot spot. Photo from NOAA.
abb22878_ch06_143-173.indd 168abb22878_ch06_143-173.indd 168 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 169
Lake
Ground surface bulging up
Lake
(a) (b)
(c) (d)
Figure 6.40 Stages in the formation of a giant continental caldera. (a) A rising mass of magma forms a low-density cap rich in SiO 2 and gases, causing the ground surface to bulge upward. (b) Plinian eruptions begin from circular fractures surrounding the bulge. (c) Magma pours out in pyroclastic flows of tremendous volume, causing the ground surface to sink into a giant caldera. (d) Removal of magma decreases the crustal pressure, allowing new magma to rise and cause the caldera floor to bulge up.
Figure 6.41 View over the town of Mammoth Lakes to tree-covered hills in Long Valley. The hills compose a resurgent dome up to 500 m (1,600 ft) high above the caldera floor. Photo from US Geological Survey.
abb22878_ch06_143-173.indd 169abb22878_ch06_143-173.indd 169 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
170 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Hot Spots Hot spots are shallow hot rock masses/magmas or plumes of slowly rising mantle rock that create volcanism on Earth’s surface. The temperature of the rising rock is hotter than the surrounding rock by about 300°C (570°F) in the plume center and only 100°C (212°F) along the outer margin of the plume head. But this temperature difference lowers viscosity enough to start the rise toward the sur- face. Most hot spots are visualized as rising plumes that operate for about 100 million years.
Hot spots do not move as much as tectonic plates and are used as reference points to help chart plate movements (see figure 2.21). They occur under the oceans and under the continents (e.g., Yellowstone), in the center of plates (e.g., Hawaii), and as part of spreading centers (e.g., Iceland). In the 1970s, a survey was made of hot spots that create elevated volcanic domes with diameters greater than 200 km (125 mi). The survey counted 122 hot spots active in the past 10 million years ( figure 6.42 ): 53 under ocean basins and 69 under continents.
The largest number of hot spots lies beneath the African plate. The drifting of Africa has been slowed by its collision with Eurasia during the last 30 million years. The slowed African plate may be acting like a thermal blanket concentrating the mantle heat beneath it. With Africa effectively stopped from making large horizontal movements, the westward movement of South America has dou- bled, and the mid-Atlantic Ocean spreading center is moving
In Greater Depth
Reunion
Yellowstone
Iceland
Azores
Canary lslands
Cape Verde
Ascension St.
Helena
Tristan da Cunha
Galapagos
Easter Island
Marquesas
Hawaii Afar
Figure 6.42 Hot spots active in the past 10 million years. Antarctica is not shown but lies above 11 hot spots, raising ques- tions about the effects of melting massive volumes of ice.
westward also, leaving some hot spots behind, as at Tristan da Cunha and St. Helena ( figure 6.42 ).
The explosiveness of volcanic eruptions above hot spots varies. They are relatively peaceful above oceanic hot spots, such as Hawaii, where low-volatile, low-viscosity, large-volume magma flows easily, analogous to spreading-center volcanism, and builds shield volcanoes (see figure 6.24 ). The Hawaiian hot spot is about 80 km (50 mi) in diameter, as defined by earthquake hypocenters at 60 km (37 mi) depth (see figure 6.25 ).
A hot spot below a spreading center means that a much greater volume of basaltic magma can erupt. For example, at Ice- land, the asthenosphere magma of the spreading process is aug- mented by deeper mantle magma to create an immense volume of basaltic rock. The combined magmas are basalt, and the eruptions are peaceful enough for the citizens of Iceland to live prosperously (see chapter 7). The mantle plume beneath Iceland is the most vigorous hot spot on Earth today. The rising plume has created crust beneath Iceland that is four to five times thicker than average.
Above continental hot spots, such as at Yellowstone National Park, the eruptions may be incredibly explosive because the rising magma breaks off and absorbs so much continental rock that it creates a volatile-rich, high-viscosity, very-large-volume magma. The mention of a big volcanic eruption may bring to mind a tall moun- tain emitting a powerful explosion, but the really big eruptions emit so much magma that they leave a hole bigger than a mountain, a giant caldera that can be 100 km (more than 60 mi) long.
abb22878_ch06_143-173.indd 170abb22878_ch06_143-173.indd 170 14/12/12 1:58 PM14/12/12 1:58 PM
G A R R E T T , M E G A N 1 3 2 4 T S
The Three Vs of Volcanology: Viscosity, Volatiles, Volume 171
A Classic Disaster Santorini and the Lost Continent of Atlantis What were the effects of the eruption of Santorini on the Mediter- ranean world? So huge that they may be the basis of the Atlantis myth.
Akrotiri was an important city, a part of the Minoan civilization based in Crete. The Minoans created an advanced civilization. In 1628 BCE , Akrotiri had three-story houses; paved streets with stone- lined sewers beneath them; advanced ceramic and jewelry work; regular trade with the Minoans’ less-advanced neighbors in Cyprus, Syria, Egypt, and Greece; and colorful wall frescoes that depicted their wealthy and comfortable life ( figure 6.43 ). In short, these
Minoans had a higher standard of living than many people in this part of the world today, more than 3,600 years later.
The dramatic collapse of this piece of the Minoan civilization must have made an indelible impression on the people of that time. In fact, this may be the event passed down to us by Plato as the disappearance of the island empire of Atlantis, which after violent earthquakes and great floods “in a single day and night disap- peared beneath the sea.” Plato lived in Greece from 427 to 347 BCE . He told the tale in the dialogues of Critias, the historian, who recounted the visit of Solon to Egypt, where he learned the account of Atlantis from the Egyptian priests in their oral histories. About 1,200 years after the event, Plato wrote a reasonably good descrip- tion of a caldera-forming collapse with attendant earthquakes, floods (steam surges or tsunami), and a landmass sinking below the sea in a day and a night.
The eruption and caldera-forming collapse into the sea at Santorini seem similar to the events at Krakatau 3,500 years later, except the Santorini event was bigger. The Santorini eruption is estimated to have blown out more than 40 km 3 (10 mi 3 ) of rhyolitic magma; Krakatau blew out 18 km 3 (4 mi 3 ). Krakatau sent out ocean waves 35 m (115 ft) high; Santorini must have done as much. The Aegean Sea region is one of the most island-rich areas on Earth. Tsunami in this region must have had a devastating effect on coastal towns and people, as well as leaving profound impressions on survivors, who passed these memories down to succeeding generations. The tales of Plato, the excavations by archaeologists, and the reconstructions by volcanologists all point to a remarkably consistent story.
This overpowering eruption was addressed by Loren Eiseley in his poem Knossus.
They died in one night, the pillars of the palace buckling, great stones cast down, the galleys beached on the shore, ruin and ashes assailing men from the sky. Thera, the burst throat of the world, coughing fire and brimstone there to the north, its voice like the bellowing of a loosed god long propitiated to no purpose. We have known it in our own lives— the fear of the moving atoms, but these people endured the actual megaton explosion, and their remnants faded from history, while the timeless, practical Egyptians regretted a small loss of trade. Civilizations die as men die, by accident then.
Figure 6.43 View of a portion of Knossos, on the island of Crete, the center of the Minoan civilization. These buildings were hit hard by the Santorini eruption. Photo by Pat Abbott.
abb22878_ch06_143-173.indd 171abb22878_ch06_143-173.indd 171 14/12/12 5:39 PM14/12/12 5:39 PM
G A R R E T T , M E G A N 1 3 2 4 T S
172 Chapter 6 Volcanic Eruptions: Plate Tectonics and Magmas
Summary Some of Earth’s internal heat causes rock to rise via con- vection and then to melt near the surface and erupt as volcanoes. Spreading centers provide such ideal settings for volcanism that 80% of all extruded magma occurs there. Plates pull apart, and magma rises up the fractures with relatively peaceful eruptions. Subduction-zone erup- tions involve magma contaminated by incorporated crustal rock, yielding high-viscosity, gas-rich magma that erupts explosively. Transform faults and continent- continent collision zones have little or no volcanism asso- ciated with them.
Hot rock at depth rises buoyantly. This hot rock may melt near the surface and become magma due to increased temperature, decreased pressure, and/or increased water content. Most magma is produced as pressure is lowered on rising hot rock via decompression melting or an increase in its water content. When magma nears the surface, gases come out of solution and help cause volcanic eruption. Whether magma erupts peacefully or explosively depends on magma types. Eruption styles and volcanic landforms can be understood via the three Vs of volcanology— viscosity, volatiles, volume. Beneath the ocean basins, magmas are basaltic in composition with low contents of SiO 2 , low weight % content of water, and high temperatures, produc- ing low viscosity, easy escape of volatiles (gases), and peace- ful eruptions. Beneath continents, rising basaltic magmas are contaminated by melting continental-crust rocks, thus altering magma compositions. The resultant andesitic- to-rhyolitic magmas have high contents of SiO 2 , high weight % content of water, and relatively low temperatures, produc- ing high viscosity, difficult escape for volatiles, and explosive eruptions.
When magma reaches the surface and gas escapes eas- ily, lava flows result. Low-viscosity lava flows may build shield volcanoes much wider than they are tall, as are found in Hawaii, for example. If gas percentage is high and the gases are trapped in magma, explosions result, blasting pyroclastic debris into the air. A scoria cone may be built around a volcanic vent by the settling of pyroclastic debris (e.g., Paricutin). Tall symmetrical volcanic peaks are usually stratovolcanoes built of alternations of lava and pyroclastic material (e.g., Vesuvius).
The volcanic explosivity index (VEI) measures the size of volcanic eruptions on a scale of 0 to 8. Between the years 1500 and 2011, one VEI 7 eruption occurred (Tambora, 1815), along with five VEI 6 events (e.g., Krakatau, 1883; Pinatubo, 1991).
Calderas form when roofs collapse into partially emptied magma chambers. This can occur when a strato- volcano is too weak to stand and its peak collapses down- ward (e.g., Crater Lake, Oregon). If the peak falls into the
ocean, major tsunami can result (e.g., Santorini, 1628 bce ; Krakatau, 1883). The biggest explosive eruptions occur on continents, where collapses may be bigger than mountains at resurgent calderas (e.g., Yellowstone).
Terms to Remember aa 154 andesite 140 basalt 158 base surge 160 caldera 163 cinder cone 160 composite volcano 160 crater 160 crystallization 149 decompression melting 152 flood basalt 159 geyser 156 lahar 162 lava dome 162 mineral 149 obsidian 154 pahoehoe 154 Plinian eruption 160
plutonic rock 149 pore 156 pumice 154 pyroclastic 154 pyroclastic flow 162 resurgent caldera 164 resurgent dome 167 rhyolite 149 rock 149 scoria 154 scoria cone 160 shield volcano stratovolcano 160 ultra-Plinian 164 viscosity 148 volatile 150 volcanic rock 149
Questions for Review 1. Mount Vesuvius is one of the world’s most active volcanoes,
yet it has quiet intervals lasting how long? Compare the times between major eruptions to a human life span.
2. Sketch a map of an idealized tectonic plate and evaluate the volcanic hazards along each type of plate edge.
3. What percentage of magma erupted each year comes out at spreading centers? At subduction zones? At hot spots?
4. What changes in temperature, pressure, and water content cause hot rock to melt? What are the two most relevant melting agents?
5. What common elements combine to form most igneous rocks?
6. What minerals combine to form most igneous rocks? 7. Contrast the differences between basaltic and rhyolitic
magma in terms of SiO 2 percentage, weight % water content, temperature, viscosity, and mode of gas escape.
8. How does explosiveness vary between magmas that have a low versus a high weight % content of water?
9. If gas escapes easily from a magma, will the eruption be peaceful or explosive? If gas cannot escape easily, will the eruption be peaceful or explosive?
10. Volcanoes in the ocean tend to erupt peacefully, whereas volcanoes on continents tend to erupt explosively. What explains the differences?
11. Why do volcanoes above subduction zones erupt more explosively than volcanoes at spreading centers?
abb22878_ch06_143-173.indd 172abb22878_ch06_143-173.indd 172 14/12/12 1:59 PM14/12/12 1:59 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Questions for Further Thought 173
20. How might a volcanic eruption create its own weather? 21. Diagram and explain the sequence of events leading to a
geyser eruption. Include temperature and pressure changes in your answer.
Questions for Further Thought 1. Why do people keep returning to a volcano, such as
Vesuvius, and building new cities? 2. Evaluate this message: Plate boundaries are bad news. 3. Would you rather watch a volcano erupt in Hawaii or
Washington? Why? 4. List the beneficial aspects of volcanoes and volcanism. 5. What evidence suggests that the eruption of the volcano
Santorini led to the enduring tale of the lost civilization of Atlantis? British Columbia—Queen Charlotte Island
12. What determines whether volcanic activity will be a lava flow or a pyroclastic eruption?
13. Which magma will make a better lava flow—basalt or rhyolite?
14. Draw a cross-section illustrating a Plinian eruption. 15. Explain the factors controlling the volcanic explosivity
index (VEI). 16. Play the three Vs game. Pick various low, medium, and
high values for viscosity, volatiles, and volume, and then describe the resultant eruption styles and volcanic landforms.
17. Draw a cross-section showing the difference between a shield volcano and a stratovolcano.
18. Draw a cross-section and describe the collapse of an oceanic volcano, such as Krakatau. What usually is the biggest killer in this process?
19. Explain the eruptive behavior of a hot spot–fed volcano on a continent.
abb22878_ch06_143-173.indd 173abb22878_ch06_143-173.indd 173 14/12/12 1:59 PM14/12/12 1:59 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcano Case Histories: Killer Events 7
CHAPTER
View over the harbor city of Catania, Sicily, toward Mount Etna. The city has been buried seven times by lavas from Etna. Each time, the city has been rebuilt on top of its predecessor to await its turn to be buried. Photo by Pat Abbott.
LEARNING OUTCOMES Active volcanoes are natural hazards, but understanding their processes helps us coexist with them. After studying this chapter, you should:
• recognize that humans can live successfully on oceanic hot-spot and spreading-center volcanoes.
• understand why subduction-zone volcanoes are so deadly.
• be able to explain the sequence of events in a catastrophic volcanic eruption as at Mount Saint Helens in 1980.
• be familiar with the leading causes of death by volcano.
• be able to explain pyroclastic flows and lahars.
• know the signs of impending volcanic eruption.
OUTLINE • Volcanism at Spreading Centers
• Volcanism at Subduction Zones
• Volcanic Processes and Killer Events
• VEIs of Some Killer Eruptions
• Volcano Monitoring and Warning
Past civilizations are buried in the graveyards of their own mistakes.
—Lord Ritchie-Calder , 1970, “Mortgaging the Old Homestead”
In te
r n
al E
n er
g y
abb22878_ch07_174-202.indd 174abb22878_ch07_174-202.indd 174 14/12/12 3:42 PM14/12/12 3:42 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanism at Spreading Centers 175
T he explosive eruption of Krakatau Volcano in Indo-nesia on 27 August 1883 seems to have influenced the world of art. Edvard Munch created his famous paint- ing The Scream ( figure 7.1 ) from an experience he had while walking in Oslo, Norway. Munch’s journal entry states: “All at once the sky became blood red . . . and clouds like blood and tongues of fire hung above the blue-black fjord and the city . . . and I felt alone, trembling with anxiety . . . I felt a great unending scream piercing through nature.” The Scream is thought to be his reac- tion to the skies made blood red in Europe by the Krakatau eruption.
In this chapter, we return to spreading-center volcanism at Iceland, examine subduction-caused explo- sive volcanism in the Cascade Range along the Pacific Coast of the United States and Canada, and then examine the historic record of volcano-related fatalities to understand the specific processes that kill people. Last, we look at failure and success in volcano monitor- ing and warning.
Volcanism at Spreading Centers Most of the volcanism on Earth takes place along the oceanic ridge systems where seafloor spreading occurs. Solid, but hot and ductile mantle rock rises upward into regions of lower pressure, where up to 30–40% of the rock melts and flows as basaltic magma. The worldwide rifting process releases enough magma to create 20 km 3 (less than 5 mi 3 ) of new basaltic oceanic crust each year. Virtually all of this volcanic activity takes place below sea level and is thus difficult to view. We see and are impressed by the tall and beautiful volcanic mountains on the edges of the continents, but the volume of magma they release is small compared to that of spreading centers.
ICELAND Iceland is a volcanic plateau built of basaltic lava erupted from a hot spot below the mid-Atlantic Ocean spreading center (see figure 4.4). The country is a little bit bigger than the state of Virginia; about 13% of its surface is covered by glaciers, and one-third consists of active volcanoes. During the nearly 1,000 years of human records, volcanic eruptions have occurred about every five years, on average. Most Icelandic eruptions do not cause deaths, but exceptions do occur (see famine of 1783 later in this chapter).
The most typical Icelandic eruptions are fissure eruptions, where lava pours out of long fractures up to 25 km (16 mi) long. To understand Icelandic eruptions, visualize the linear spreading center (see figure 4.5) that controls the rise of magma as it is fed upward through fractures. An Ice- landic eruption can be beautiful to watch as an elongate “curtain of fire” shoots upward with varying intensity and height. Icelandic eruptions of low-viscosity, low-volatile lava flows can be so peaceful that their movement is almost waterlike.
Lava Flows of 1973 The recent story of Iceland shows that humans can make enough adjustments to live profitably and happily next to active basaltic volcanism. The 1973 eruptions on the small island of Heimaey on the southern coast of Iceland illustrate the “peaceful” nature of these eruptions. The town of Vestmannaeyjar is built next to the premiere fishing port in Iceland. The safe harbor is itself a gift of volcanism; it was formed between ancient lava flows. On 23 January 1973, a fissure opened up only 1 km (3,300 ft) from the town of 5,300 people ( figure 7.2 ). By early July, the eruption had emitted 230 million m 3 of lava ( figure 7.3 ) and 26 million m 3 of pyroclastic material. The lava flows increased the size of the island by 20%. Gases vented during the eruptive sequence, other than water vapor, were dominantly CO 2 with lesser amounts of H 2 , CO, and CH 4 . The only fatality was a person asphyxiated in a gas-filled building.
Figure 7.1 The Scream, painted by Edvard Munch, is thought to be his reaction to the skies made blood red in Europe by the Krakatau eruption. Photo © Erich Lessing/Art Resource, NY. Art: © 2010 The Munch Museum/The Munch-Ellingsen Group/Artists Rights Society (ARS), NY.
abb22878_ch07_174-202.indd 175abb22878_ch07_174-202.indd 175 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
176 Chapter 7 Volcano Case Histories: Killer Events
controlled the flow of later lavas and even controlled the flow paths of the dense volcanic gases. To save their harbor and economic livelihood, the Icelanders sprayed seawater on the lava flows, causing rapid cooling and hardening into wall- like features that forced the lava to flow off in another direc- tion ( figure 7.4 ). This action prevented the harbor from being filled and closed. Now, with its new shape and larger size, the harbor is better than before the 1973 eruption (see figure 7.2 ).
When the eruptions stopped, the people set up a pipe system that poured water into the 100 m (330 ft) thick mass of slowly cooling lava. Return pumps were installed to bring the water, which had been heated to 91°C (196° F), back to the surface and into town, where it was used to heat buildings. Basaltic eruptions do not have to be killers. Humans and volcanoes can coexist in harmony, with luck and with some exceptions.
Volcanism at Subduction Zones Through newspapers and television, we learn of death-dealing volcanic eruptions at Galeras Volcano in Colombia, Mount Unzen in Japan, Mounts Pinatubo and Mayon in the Philippines, Mount St. Helens in Washington, and Soufriere Hills on Montserrat. These are all subduction-zone volcanoes. These stratovolcanoes have the biggest impact on humans. Many of the regions around subduction-zone volcanoes are heavily populated and feel the wrath of the eruptions. Also, because these volcanoes erupt directly into the atmosphere, they can affect climate worldwide (see chapter 12).
HEIMAEY
VESTMANNAEYJAR
Original coastline
New land
1973 eruptive fissure
Harbor
1 Mile
1 Kilometer
N
Figure 7.2 The island of Heimaey with the old coast- line shown as an orange line. The dark gray area is new land formed by the 1973 lava flow. Note that the new harbor is bigger and better protected. Data source: Williams, R. S., Jr., and Moore, J. G., “Man Against Volcano: The Eruption on Heimaey, Vestmannaeyjar, Iceland,” US Geological Survey, 1983.
Figure 7.3 An aa lava flow stopped against and between two fish-factory buildings in Vestmannaeyjar, 23 July 1973. Photo from US Geological Survey.
Figure 7.4 Seawater is being sprayed on the lava front to cool, harden, and stop it from closing off Vestmannaeyjar harbor, 4 May 1973. Photo from US Geological Survey.
The early lava flows on Heimaey began filling in the harbor and destroying about 300 buildings; pyroclastic fall- out buried another 70 buildings. But the volume of lava was not overwhelming, so the Icelanders took over. Pyroclastic material was bulldozed to create barriers that diverted and
abb22878_ch07_174-202.indd 176abb22878_ch07_174-202.indd 176 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanism at Subduction Zones 177
CASCADE RANGE, PACIFIC COAST OF UNITED STATES AND CANADA Explosive eruptions are frequent at the numerous volcanoes in the Pacific Northwest region of the United States and in British Columbia ( figure 7.5 ). The plate-tectonic process responsible for these volcanoes is identical to the cause of the region’s great earthquakes—subduction. In fact, the frequent eruptions from the Cascade Range volcanoes provide clear evidence for active subduction. The melting of part of the mantle (asthenosphere) wedge above the subducting plate is aided by water released from sediments on top of the sub- ducting plate. The rising basaltic magma partially melts overlying crustal rock as well, increasing its content of SiO 2 and water. Much of the magma changes its composition to andesite or rhyolite and increases its viscosity as it rises ( figure 7.6 ). Some collects in great pods and cools under- ground, forming plutonic rocks, but some erupts explosively at the surface.
How often do major eruptions occur? An example was documented in a 1975 study of Mount St. Helens by Dwight Crandell and colleagues. Their report stated that the latest large, volcanic mountain had formed in the last 2,500 years. Since then, Mount St. Helens has experienced major erup- tions every century or two and has never been free from major volcanism for longer than 500 years. Figure 7.7 shows
Portland Mount Hood
Mount Jefferson
Mount Thielsen
Klamath Falls
Medicine Lake
Lassen Peak
Mount Shasta
Mount McLoughlin
Crater Lake
Newberry Volcano
British Columbia
Washington
Mount RainierMount St. Helens Mount Adams
Simcoe Volcanic field
Seattle
Glacier Peak
Mount Baker
Vancouver
North American plate
Pacific plate
Juan de Fuca plate
Gorda plate
Mount Garibaldi
Oregon
California
Three Sisters
Eugene
Nevada
0
0
100
60
200km
125mi
Explorer plate
Figure 7.5 Plate-tectonic map of Cascade Range volca- noes. Volcanoes are subparallel to the subduction zone and spaced somewhat regularly.
Pacific Ocean
Cascade Range
Mixing of magma
Continental crust
Rhyolitic magma
Continental lithospherePartial melting of
continental crust
Basaltic magma
Water aids partial melting of asthenosphere
Water added from subducting crust and sediments
Oceanic crust (basalt)
Asthenosphere
Oceanic lithosphere
0 100km
60mi0
Andesitic magma
Figure 7.6 Subduction-zone volcano “factory.” Basaltic magma forms in the upper asthenosphere where subducted water aids partial melting of oceanic crust and asthenosphere. Rising magma partially melts some continental crust, forming water-rich andesitic to rhyolitic magmas that erupt explosively.
abb22878_ch07_174-202.indd 177abb22878_ch07_174-202.indd 177 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
178 Chapter 7 Volcano Case Histories: Killer Events
the age distribution and eruption history along the line of Cascade Range volcanoes. Basically, the volcanoes are all the same age; they sit above subducting plates and are active now. Volcanoes built above hot spots also line up—for example, Hawaii (see figure 2.23) and Yellowstone (see figure 6.38). In contrast to subduction-zone volcanoes, the ages along lines of hot-spot volcanoes range from young to old in orderly progressions.
How are prehistoric eruptions documented? The process is the same as that used to work out dates of prehistoric earth- quakes. The slopes near a volcano reveal the remains of trees knocked down by volcanic blasts ( figure 7.8 ). These trees may be buried by volcanic ash, incorporated in lahars, or otherwise preserved. Radiocarbon determinations of the dates when trees died also tell the dates of the volcanic eruptions that killed them ( figure 7.9 ). The 1975 report stated, “Although dormant since 1857, St. Helens will erupt again, perhaps before the end of this century.” The geologic analysis was prophetic ( figure 7.10 ).
Mount St. Helens, Washington, 1980 In late March 1980, Mount St. Helens awoke from a 123-year- long slumber. Dozens of magnitude 3 earthquakes occurred each day as magma pushed its way toward the surface. On 27 March, small explosions began as groundwater and magma came in contact. The spectacle of an erupting volcano was a
tremendous lure for sightseers. People flocked to Mount St. Helens. The weekend traffic was so jammed that it reminded folks of rush hour in big cities. But this was an explosive giant just warming up its act, and all nearby life was in grave danger. The governor of Washington ordered block- ades placed across the roads to Mount St. Helens to keep people away. Her action was unpopular. Then, at 8:32 a.m. on 18 May 1980, the volcano blew off the top 400 m (1,313 ft)
Meager Mountain
Mount Garibaldi
Mount Baker
Glacier Peak
Mount Rainier
Mount St. Helens
Mount Adams
Mount Hood
Mount Jefferson
Three Sisters
Newberry Volcano
Crater Lake
Mount McLoughlin
Medicine Lake Volcano
Mount Shasta
Lassen Peak
BC
WA
OR
CA
P ac
ifi c
O ce
an
4,000 2,000 200 Years ago Present
Figure 7.7 Eruption histories of Cascade Range volcanoes during the last 4,000 years.
Figure 7.8 What was once a mature forest is now a field of fallen trees pointing in the direction traveled by the volcanic blast from Mount St. Helens on 18 May 1980.
abb22878_ch07_174-202.indd 178abb22878_ch07_174-202.indd 178 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanism at Subduction Zones 179
Lahar
Lahar
Ash deposits
Buried tree stumps
Buried logs
Figure 7.9 Schematic cross-section of a volcano and some of its eruptive deposits. Radiocarbon dates on buried wood tell when trees died—that is, when the volcano erupted.
of its cone during a spectacular blast that generated about 100 times the power of all U.S. electric-power plants combined. Most of the 62 people killed had found ways around the bar- ricades to get a better view of an erupting volcano. A look at the eruptive sequence provides a good example of how an explosive volcano does its thing ( figure 7.11 ).
First, Mount St. Helens achieved its beautiful conical shape during the mid-1800s ( figure 7.11 a ). In 1843, a SiO 2 - rich lava dome grew at the volcano peak. In 1857, andesitic lava flows cooled high on the slopes. But these events also set up discontinuities, or weaknesses, within the volcanic cone.
Second, in 1980, rising magma began changing the shape of the volcano ( figure 7.11 b ). Earthquake hypocenters were abundant at 1 to 3 km (0.6 to 2 mi) depth. The seisms were recording the injection and pooling of magma. With magma forcing its way upward, the northern side of the volcano began rising. The increasing volume of magma also caused the groundwater body to expand its volume. The effect on the volcano was dramatic. By 12 April, a 2 km 2 (1.2 mi 2 ) area on the north flank had risen upward and out- ward by 100 m (330 ft). This unstable situation grew worse as the “mega-blister” kept growing about 1.5 m (5 ft) per day.
Third, at 8:32 a.m. on 18 May 1980, the bulge failed. With magma injecting into the bulge from below and gravity pulling from outside, the huge mass of the bulge, with its weak strength, failed and pulled away as an avalanche. The shaking ground was recorded as a magnitude 5.1 earthquake. The avalanche material was 2.5 km 3 of the north side of the mountain; it fell away at speeds up to 250 km/hr (150 mph) ( figure 7.11 c ). The avalanche was a roiling mass of frag- mented rock that once was the mountaintop and side, combined with ice blocks, snow, magma, soil, and broken trees; the internal temperature of the mass was about 100°C (212°F). Part of the avalanche slammed into Spirit Lake,
causing waves 200 m (650 ft) high. Another part overrode a 360 m (1,180 ft) high ridge that lay 8 km (5 mi) to the north; then it turned and moved 23 km (14 mi) down the north fork of the Toutle River ( figure 7.12 ). The resulting deposit was a chaotic mixture of broken rocks and loose debris that aver- aged 45 m (150 ft) in thickness and had a hummocky surface relief of 20 m (65 ft). Only a short time earlier, this material had been the top of the mountain. At the same time as the avalanche occurred, lahars were forming and flowing down the river valleys as rock particles mixed with water derived from melting snow and ice, from Spirit Lake, and from within the avalanche. These slurries continued to form and flow for many hours after the eruption began. Lahars moved long distances at speeds up to 40 km/hr (25 mph), carrying huge boulders and flowing with a consistency like wet concrete.
Fourth, as the landslide began to pull away, the dramatic drop in pressure on the gaseous magma and superheated groundwater caused a stupendous blast ( figure 7.11 d ). The blast and surge roared outward at speeds up to 400 km/hr (250 mph). The blast overtook and passed the fast-moving avalanche, racing over four major ridges and scorching an area of 550 km 2 (210 mi 2 ) with 0.18 km 3 of volcanic rock fragments and swirling gases at about 300°C (572°F) ( figure 7.12 ). The blast was a pyroclastic flow. It was denser than air, flowing along the ground as a dark cloud, with turbulent volcanic gases keeping solid rock fragments, magma bits, and splintered trees in suspension; it behaved as a very low-viscosity fluid.
Fifth, the big blast opened up the throat of the volcano, exposing an effervescing magma body. Rapidly escaping gases blew upward, carrying small pieces of magma to heights greater than 20 km (12 mi) during the Plinian phase, which lasted about nine hours ( figure 7.11 e ). The boiling gases carried about 1 km 3 (0.24 mi 3 ) of volcanic ash up and
abb22878_ch07_174-202.indd 179abb22878_ch07_174-202.indd 179 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
180 Chapter 7 Volcano Case Histories: Killer Events
Figure 7.10 Mount St. Helens, Washington. (a) Before: View to the northeast of the beautiful cone of Mount St. Helens on 25 August 1974. Mount Rainier is in the distance. (b) After: Same view on 24 August 1980, after the volcano had blown off its top 400 m (1,313 ft). Photos by John S. Shelton.
(b)
(a)
abb22878_ch07_174-202.indd 180abb22878_ch07_174-202.indd 180 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanism at Subduction Zones 181
1857 Andesitic lava flow
Crater 1843 pyroclastic flow and later SiO2-rich dome
SiO2-rich dome
Toutle River Valley
Toutle River Valley
Toutle River Valley
Cracks
Bulge collapsing
Landslide/ avalanche
Blast
Pyroclastic flows
Ash
Mudflows
(a) Preeruption
(e) Vertical eruption (Plinian)
(f) Dome building
Toutle River Valley
Toutle River Valley
(b) Bulge
(c) Landslide/avalanche
Toutle River Valley
(d) Blast
Ash
Bulge due to pressure from expanding gases and magma
Plane of failure
Figure 7.11 Eruptive sequence (VEI = 5) of Mount St. Helens in 1980. (a) The symmetrical volcanic cone was shaped in 1843 and 1857. (b) In late March, rising magma and expanding gases caused a growing bulge on the northern side. (c) At 8:32 a.m. on 18 May 1980, the bulge failed in a massive landslide/debris avalanche recorded as a magnitude 5.1 earthquake. (d) The landslide released pressure on the near-surface body of magma, causing an instantaneous blast of fragmented rock and magma. (e) The “throat” of the volcano was now clear, and the vertical eruption of gases and small blobs of magma shot up to heights of more than 20 km (12 mi) for nine hours. (f) Today, the mountain is slowly rebuilding a volcanic dome of low–water content, SiO 2 -rich magma.
away. About 0.25 km 3 of ash was blown across the United States at different heights by various wind systems. Another 0.25 km 3 formed pyroclastic flows by either spilling out of the volcano or falling down from the eruption cloud ( figure 7.13 ). These pyroclastic flows had temperatures of 300° to 370°C (570° to 700°F) and moved at speeds up to 100 km/hr (more than 60 mph).
Today, the volcano is slowly repairing the damage done to its once-symmetrical cone as it builds an SiO 2 -rich lava dome ( figures 7.11 f and 7.14 ). The magma building the lava dome has not erupted explosively, probably because it lost most of its volatiles during the big eruption on 18 May 1980.
Mount St. Helens looks very different these days ( figure 7.15 ). Gone are the mountaintop, snowfields, forests, and lakes. The once tree-lined river valleys are clogged with volcanic debris ( figure 7.16 ). But recovery is progressing well. Bacteria have eaten the sludge from dirty lakes, leaving pure water that has been stocked with trout. Plants have sprouted anew in devastated ground, and animals have returned to feed on them and on each other. Life is erasing the effects of the volcanic events.
Were the explosive events at Mount St. Helens a rare occurrence? Are similar events likely at other Cascade Range volcanoes in our lifetimes?
abb22878_ch07_174-202.indd 181abb22878_ch07_174-202.indd 181 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
182 Chapter 7 Volcano Case Histories: Killer Events
Green River
Coldwater
Creek
Sp iri
t La
ke
C le
ar w
at er
C re
ek
South Fork Toutle River
P ine C
reek
M uddy R
iver
Outline of St. Helens crater Pyroclastic flow deposit Mudflow deposits Lateral blast zone Avalanche deposit
0
0
4 8 km
5 mi
North Fork Toutle River
Figure 7.13 High-temperature pyroclastic flow rolling down the side of Mount St. Helens, 7 August 1980. Photo from US Geological Survey.
Figure 7.12 Map of materials dumped in the 18 May 1980 eruption of Mount St. Helens. Avalanche deposit was from the initial landslide. A lateral blast followed immediately. Then pyroclastic flows spilled out of the exposed magma body. Through it all, the superheated groundwater, plus melting snow and ice, fluidized sediments on the steep slopes as lahars (mudflows) that ran down the valleys. Source: R. Tilling, “Eruptions of Mount St. Helens: Past, Present and Future,” 1984, US Geological Survey.
Figure 7.14 Lava dome of high-viscosity, low-volatile magma growing in the central magma pipe of Mount St. Helens since its big eruption in 1980. Photo © PhotoLink/Photodisc/Getty Images RF.
abb22878_ch07_174-202.indd 182abb22878_ch07_174-202.indd 182 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanism at Subduction Zones 183
Figure 7.15 View of the devastated northeastern side of Mount St. Helens, 20 August 1980. Mount Hood is in the background. Photo by John S. Shelton.
Figure 7.16 The Toutle River, choked with eruption debris. Photo by John S. Shelton
abb22878_ch07_174-202.indd 183abb22878_ch07_174-202.indd 183 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
184 Chapter 7 Volcano Case Histories: Killer Events
Lassen Peak, California, 1914–1917 Lassen Peak is not the typical volcano; rather, it is an unusu- ally large (about 1 mi 3 ) lava dome of SiO 2 -rich volcanic rock analogous to that growing in Mount St. Helens today (see figure 7.14 ). Lava domes form when magma is too poor in volatiles and too viscous to flow away, so instead it oozes upward as a conduit-plugging mass (see figure 6.32).
Lassen Peak awakened in May 1914 with numerous eruptions, culminating on 18 July 1914 with a major episode that sent up an ash cloud more than 3,350 m (11,000 ft) high. Small-scale volcanic activity continued, but large events did not resume until May 1915. On 16–18 May 1915, the 300 m (1,000 ft) wide crater overfilled with water-deficient, sticky magma that stood higher than the rim. The magma was too viscous to flow over the lip, so instead, red-hot blocks broke off and rolled downslope. Meanwhile, to the east, the melting snow combined with rocky debris to set in motion a massive lahar that flowed outward 50 km (30 mi). On 19 May, on the north slope, the side of Lassen Peak split, and a pyroclastic flow blasted forth as a mixture of superhot gases, fragmental rock debris, trees, and water, devastating a triangular-shaped area 6.5 km (4 mi) long and 1.6 km (1 mi) wide ( figure 7.17 ). Volcanic activity continued with more lahars and pyroclastic flows, and on 22 May, a broad mushroom cloud of ash was blasted 8 km (5 mi) high. Lassen remained relatively peace- ful through 1916, but May and June 1917 brought renewed activity.
In three of four years, the month of May saw the start of extensive volcanic activity. Was this a coincidence? Maybe, but it is possible that as water from the melting snow sank and was heated underground, its volume expansion helped fracture Lassen Peak and reduce internal pressure enough to begin the eruptions. In the nonvolcanic year of 1916, Lassen Peak was too hot for snow to accumulate.
In the 20th century, two Cascade Range volcanoes underwent similar eruptions with sideward-directed blasts, pyroclastic flows, far-reaching volcanic mudflows (lahars), and great vertical eruptions of ash (Plinian phase). Luckily, each of these eruption sequences took place in sparsely inhabited areas. What are the prospects for similar eruptions near towns and cities?
Mount Shasta, California Mount Shasta, at 4,318 m (14,162 ft) elevation, is the second tallest of the Cascade Range volcanoes ( figure 7.18 ). The third highest is Shastina (3,759 m, or 12,330 ft), perched on its shoulder. The combined mountain mass is particularly impressive, standing over 3,000 m (10,000 ft) higher than its surroundings and visible from more than 160 km (100 mi) away. Mount Shasta is an active volcano, erupting 11 times in the last 3,400 years, including at least 3 times in the last 750 years. Its last eruption was probably in 1786.
The Mount Shasta area is a beautiful place to live, and the towns along the volcano base are growing. But how wise
Figure 7.17 View of the north side of Lassen Peak, devastated by the 19 May 1915 eruption. Photo by John S. Shelton.
abb22878_ch07_174-202.indd 184abb22878_ch07_174-202.indd 184 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanism at Subduction Zones 185
Figure 7.18 View from the north to Mount Shasta and Shastina. Note the network of roads being used to develop towns on top of lava flows, lahars, and debris avalanche deposits. Photo by John S. Shelton.
Most likely
Less likely
Lesser likelihood
Gra ve
l C r.
Mount ShastaPilgrim Cr.
Ash Cr.
Shastina Lake
S ha
st a
R iv
er
W hitney C
r.
Shastina
Weed
Diller Canyo n
Black Butte
C as
ca de
G ul
ch
Mount Shasta (town)
Highway 89 Sacramento
River
Siskiyou Lake
Dunsmuir McCloud River
McCloud
Mud C r.
Interstate 5
Edgewood
H ig
hw ay
9 7
0 5 mi
8 km0
M ou
nt ai
n H ous
e C r.
Brew er Cr.
N
Figure 7.20 Map of the Mount Shasta–Shastina area showing the most likely paths for lahars. These volcanic mud- flows tend to occupy the same river bottom flatlands where towns are built. Source: D. R. Crandell and D. R. Nichols, “Volcanic Hazards at Mt. Shasta,” 1989, US Geological Survey.
The Whaleback
Ash Creek Butte
Dunsmuir
Most frequent
Less frequent
Lesser frequency 0 5 10 mi
0 15 km
Edgewood
Weed Shastina Mount Shasta
Black Butte
Mount Shasta (town)
Interstate 5
Lake Shastina
Figure 7.19 Map of the Mount Shasta–Shastina region, showing the areas most susceptible to lateral blasts and pyro- clastic flows. Note the growing towns within the danger zones. Source: D. R. Crandell and D. R. Nichols, “Volcanic Hazards at Mt. Shasta,” 1989, US Geological Survey.
is this? The lower slopes of Mount Shasta are broad and smooth, allowing pyroclastic flows to spread widely as they move down the volcano flank ( figure 7.19 ). Lahars are more prone to flow through valleys, and towns lie there ( figure 7.20 ). The rock record gives further reason to pause and consider whether or not to build here. Figure 7.21 shows the distribution of a 300,000-year-old avalanche deposit that extends 43 km (27 mi) out from the volcano base. This catastrophic event deposited eight times more debris than Mount St. Helens did in 1980. This jumbled mass near Mount Shasta is the foundation for three towns and one large reservoir.
Would it be advisable to draw park boundaries around the hazardous Cascade Range volcanoes and not allow towns to be built there? Volcanologists Dwight Crandell, Donal Mullineaux, and Meyer Rubin point out that
The potential risk from future eruptions may be low in rela- tion to the lifetime of a person or to the life expectancy of a specific building or other structure. But when dwelling places and other land uses are established, they tend to persist for centuries or even millennia.
abb22878_ch07_174-202.indd 185abb22878_ch07_174-202.indd 185 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
186 Chapter 7 Volcano Case Histories: Killer Events
Volcanic Processess and Killer Events Volcanoes can kill in numerous ways ( figure 7.22 ). They can burn you with a pyroclastic flow, slam and suffocate you with a lahar, batter and drown you with tsunami, poison you with gas, hit you with a pyroclastic bomb, fry you with a lava flow, or kill you with indirect events such as famine.
THE HISTORIC RECORD OF VOLCANO FATALITIES Volcanoes operate all around the world. How many people do they kill? Which volcanic processes claim the most lives? The lack of written records for some time intervals and in some parts of the world makes these questions difficult to answer. Volcanologists Tom Simkin, Lee Siebert, and Russell Blong have studied the questions and given approximate answers. About 275,000 people have been killed by volcanic action during the past 500 years ( figure 7.23 ). A dozen or so volcanic processes have done the killing ( table 7.1 ). We will now individually examine some of the killer processes. As
Klamath River Gorge
Shasta River Gorge
Montague
C ascade R
ange
Yreka
S hasta R
iver
Dam
97
Mount Shasta
Shastina
Weed
Edgewood
Mount Shasta (town)
K lam
ath M ountains
0
0
5 10 km
6 mi
5
Lake Shastina
Figure 7.21 Map of a 300,000-year-old debris avalanche deposit at the base of Mount Shasta. The amount of material is eight times greater than the amount erupted at Mount St. Helens in 1980. It forms the foundation for three towns and one reservoir. See an avalanche deposit in the front center of figure 7.18 .
we do so, note that several of these each resulted in thousands of deaths ( figure 7.23 ).
PYROCLASTIC ERUPTIONS Explosive volcanic eruptions shatter magma into pieces by gas bubble growth and blast the fragmented magma as pyro- clasts up to the surface and into the atmosphere. The pyro- clasts may be brought back to the ground as pyroclastic falls, pyroclastic flows, and pyroclastic surges.
Pyroclastic Falls During an explosive eruption, airborne pyroclasts fall down on the landscape with particles ranging in size from ash to bombs to huge blocks. This pyroclastic fall is similar to can- non bombardments during war, and it results in about 2% of volcanic deaths.
Pyroclastic Flows Few experiences on Earth are as frightening as having a super-hot, turbulent cloud of ash, gas, and air come rolling toward you at high speed. History records numerous instances of pyroclastic flows killing thousands of people at each event.
A pyroclastic flow is an overwhelming mixture of hot hunks of magma, volcanic ash, volcanic gas, and mixed-in air that flows downslope at speeds greater than 10 m/sec (22 mph) and may exceed 100 m/sec (225 mph). Pyroclastic flows derive their energy from the volcanic eruption, gas expansion within the flowing mass, and the pull of gravity. Temperatures of 350°C (660°F) were measured inside the volcanic ash cloud at Mount Unzen, Japan, in 1992.
Pyroclastic flows are responsible for 29% of volcanic deaths; they are the deadliest volcanic process ( table 7.1 ). Pyroclastic flows begin in a variety of ways ( figure 7.24 ).
Dome Collapse A growing lava dome provides a unique combination of steady magma supply and the upward lift of unstable, overhanging topography. Big hunks of lava dome frequently break off and create pyroclastic flows ( figure 7.24 a ). At Mount Unzen in Japan, between 1991 and 1994, more than 7,000 dome collapses were recorded.
In May 1991, the lava dome in Mount Unzen began a growth spurt that attracted international attention. As the unstable lava dome grew and towered 90 m (300 ft) above the crater rim, 15,000 residents were evacuated from villages and tea plantations around the mountain’s base. As residents left, journalists and volcanologists arrived to record the numerous collapses of 200 to 300 ft high masses from the lava dome and watch the debris run downslope as glowing pyroclastic flows. At 4:09 p.m., on 3 June 1991, a much larger than usual mass fell off the lava dome and rolled downslope at about 60 mph, killing 44 observers, including the famed French volcano photographers, Maurice and Katya Krafft. All the deaths occurred in previously evacu- ated areas.
abb22878_ch07_174-202.indd 186abb22878_ch07_174-202.indd 186 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanic Processess and Killer Events 187
GroundGroundwateater
Ash fall
Acid rain
Lava flow
Lahar (mudflow)
Pyroclastic flow
Lava dome collapse
Eruption cloud
Lava dome
Bombs
Prevailing wind
Gases
Groundwater
Magma
Crack
Eruption column
Pyroclastic flow
Debris avalanche
Figure 7.22 Volcanoes operate many life-threatening natural processes. Source: US Geological Survey Fact Sheet 002–97 (1997).
1500
300,000
250,000
200,000
150,000
100,000
C um
ul at
iv e
fa ta
lit ie
s
50,000
0 1600 1700
Unzen 1792
Kelut 1586
Krakatau 1883
Pelée 1902
Ruiz 1985
Tambora 1815
Laki 1783
Year 1800 1900 2000
Figure 7.23 Cumulative fatalities from volcanoes during 500 years, 1500 to 2000. Data Source: T. Simkin, L. Siebert, R. Blong, Science 291:255 (2001).
abb22878_ch07_174-202.indd 187abb22878_ch07_174-202.indd 187 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
188 Chapter 7 Volcano Case Histories: Killer Events
TABLE 7.1 Volcanic Causes of Deaths
275,000 Deaths
530 Volcanic Events
Pyroclastic flow 29% 15%
Tsunami 21% 5%
Lahar 15% 17%
Indirect (famine) 23% 5%
Gas 1% 4%
Lava flow �1% 4%
Pyroclastic fall (bombs) 2% 21%
Debris avalanche 2% 3%
Flood 1% 2%
Earthquake �1% 2%
Lightning �1% 1%
Unknown 7% 20%
Data source: Simkin, T., Siebert, L., and Blong, R., “Volcano fatalities” in Science 291:255, 2001.
(d) Eruption column collapse
(c) Direct blast
(b) Overspilling crater rim
(a) Dome collapse
Figure 7.24 Ways of generating pyroclastic flows: (a) dome collapse as at Mount Unzen, 1991; (b) overspilling of crater rim as at Mont Pelée, 1902–1903; (c) direct blast as at Mount St. Helens, 1980, and Mount Pinatubo, 1991; (d) eruption column collapse as at Mount Mayon, 1984.
Overspilling Crater Rim A volcano crater may have its lake turned into a boiling cauldron, or the crater may fill with magma. If the crater overfills, hot water and magma can pour over the rim and flow downslope ( figure 7.24 b ). This happened numerous times on Mont Peléee in Martinique in 1902.
Direct Blast In some eruptions, a pyroclastic flow may simply form as a direct blast from the volcano. In 1980, as a landslide moved down Mount St. Helens, the decrease in pressure on magma inside the volcano caused a tremendous direct blast (see figure 7.11 d ). The direct blast traveled 150 m/sec (335 mph) and overwhelmed everything it encountered—lakes, trees, people.
Eruption Column Collapse At its greatest power, a volcano may send its eruption column of hot pyroclastic material, hot gas, and intermixed air high up into cooler air, providing time for heat to dissipate and for pyroclasts to cool and be spread far and wide. A dangerous phase of the erup- tion can occur in those moments when less energy is fed into the eruption column and the column begins to collapse, sending clouds of hot gases, ash, and pumice flowing as ground-hugging deadly pyroclastic flows ( figure 7.27 ).
Since 1616, more than 1,500 people have been killed during 40 recorded deadly eruptions of the subduction- caused stratovolcano Mount Mayon in the Philippines. In 1984, a series of Vulcanian eruptions sent magmatic debris 10 km (6 mi) into the atmosphere several times. Partial col- lapses of the eruption column sent pyroclastic flows rolling down the mountain slope at velocities ranging from 50 to 100 km/hr (30 to 60 mph) ( figures 7.24 d and 7.27 ).
Pyroclastic Flows over Water Can a pyroclastic flow travel across a body of water to kill you? Or does the water
absorb heat from the hot, gas-rich cloud quickly enough to eliminate its ability to kill? A body of water does not elimi- nate the hazard. During the 1883 eruptions of Krakatau in Indonesia leading up to the volcano collapse, one remarkable blast on 27 August sent out a hot, gaseous pyroclastic flow that raced across the sea surface of the Sunda Straits for 40 km (25 mi) to reach the coastal province of Katimbang on Sumatra (see figure 8.21). It flowed onshore with enough heat to fatally burn more than 2,000 people.
Pyroclastic Surges Pyroclastic surges occur when more steam and less pyro- clastic material combine to produce a more-dilute, less- dense, high-velocity flow. Because of their low density, surges are not as easily controlled by topography. Some
abb22878_ch07_174-202.indd 188abb22878_ch07_174-202.indd 188 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanic Processess and Killer Events 189
A Classic Disaster
Atlantic Ocean
Martinique
Basse Pointe
Grande Rivière
Le Prècheur
Sugar mill
St Pierre
Carbet
Morne Rouge
Ajoupa Bouillon
Area scorched by 30 August 1902 nuée ardente
Area scorched by 8 May 1902 nuée ardente
Mont Pelèe
Caribbean Sea
R ivi
èr e
Bl
an ch
e
8 km0
0 1 2 3 4
4
5 mi
Figure 7.25 Map of Mont Pelée showing areas scorched by the largest pyroclastic flows of 1902.
Mont Pelée, Martinique, 1902 The Caribbean island of Martinique in the West Indies was colo- nized by the French in 1635. The tropical climate was superb for growing sugarcane to help satisfy the world’s growing appetite for the sweetener. On the north end of Martinique is a 1,350 m (4,430 ft) high volcano. The French called the volcano Pelée, mean- ing “peeled” or “bald,” to describe the bare area where volcanism had destroyed all plant life during the eruptions of 1792 and 1851. By coincidence, the pronunciation of the French word Pelée is the same as the Polynesian word Pele used in Hawaii to denote the goddess of volcanoes and fire.
In early spring of 1902, Vulcanian activity began. The crater atop Mont Pelée began filling with extremely viscous magma, displacing boiling lake waters through a V-shaped notch ( figure 7.25 ). The extraordinarily sticky magma kept plugging the crater. At times, superhot pyroclastic flows would spill out of the crater; at other times, they would blast out. By late April, it was obvious to most people that this problem might get bigger. About 700 rural folks were migrating each day into St. Pierre, a city of pictur- esque, early 17th-century buildings that normally was home to 25,000 residents. Another 300 people a day were leaving St.
Pierre, which lay only 10 km (6 mi) from Mont Pelée. At a little past noon on 5 May, a large pyroclastic flow sped down the Riviére Blanche, destroying the sugar mill and 40 people. This further increased the anxiety level in St. Pierre. But there was an election coming up on 10 May, and the governor did not want everyone scattered from the island’s largest city because that would likely change the election results. Governor Mouttet and his wife went to St. Pierre and used the militia to preserve order and halt the exodus of fleeing people. Bad decision. There was no election on 10 May anyway; all the voters, including the governor, died on 8 May ( figure 7.26 ).
On the morning of 8 May 1902, a massive volume of gas- charged, ultrasticky magma had risen to the top of the crater. At about 7:50 a.m., witnesses heard sharp blasts that sounded like thousands of cannons being fired as trapped gas bubbles exploded and shattered magma into fine pieces. This spectacu- lar pyroclastic flow moved as a red-hot avalanche of incandes- cent gases and glowing volcanic fragments (then called nuée ardente, which is French for “glowing cloud”). The mass moved as solid particles of magma suspended in gas. Its energy came from (1) the initial blast, (2) gravity, and (3) gas continuing to escape from the pieces of airborne magma, creating a “pop- corn” effect. The momentum of the flow was aided and its friction reduced by internal turbulence and air mixed into the flow as it moved downward and outward. The temperature at the crater is estimated to have been about 1,200°C (2,200°F), and the glowing cloud was still hotter than 700°C (1,300°F) when it hit St. Pierre. The coarsest and heaviest part of the pyro- clastic flow moved down the Riviére Blanche. The associated
Figure 7.26 The pyroclastic flow–charred remains of St. Pierre, May 1902. Mont Pelée is in the background. Photo by Underwood and Underwood/Library of Congress.
abb22878_ch07_174-202.indd 189abb22878_ch07_174-202.indd 189 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
190 Chapter 7 Volcano Case Histories: Killer Events
pyroclastic debris settling from the atmosphere was uncom- fortable, but the Plinian event was not enough to drive the rural farmers and villagers from their land. The next five days were calming for the residents, as only minor volcanic activity occurred. But suddenly, on 4 April, a pyroclastic surge flowed radially outward for 8 km (5 mi), overrunning nine villages and killing 2,000 people. Everyone within 8 km of the volcano, in any direction, was killed by the base surge. Following the surge, a Plinian column shot up 20 km (12 mi). On the same day, there were two more base surges and Plinian columns, but the last two base surges did not matter; everyone was already dead. In addition, the Plinian columns injected sulfur dioxide (SO 2 ) into the upper atmo- sphere, and the whole world felt the effect as global climate changed.
TSUNAMI Volcanic tsunami can be created when some of the huge amounts of energy produced during volcanic eruptions are injected into large water bodies. Volcanic processes that generate tsunami include caldera collapse into the ocean; undersea eruptions; and travel of pyroclastic flows, lahars, and debris avalanches into the sea. Volcano-generated tsunami have been responsible for 21% of volcano-caused deaths.
Caldera Collapse The collapse of Krakatau Volcano in 1883 killed more than 36,000 people. The volcanic eruptions directly killed less than 10% of the people; more than 90% of the fatalities were due to volcano-caused tsunami. The Krakatau tsunami are discussed in detail in chapter 8.
Figure 7.27 Formation of pyroclastic flows as collapses from the vertical eruption column flow downhill, Mayon Vol- cano, Philippines, 1984. Photo by Chris G. Newhall, US Geological Survey.
volcanic eruptions that involve magma and water interaction produce ground-hugging surges that may flow in all direc- tions simultaneously as ring-shaped base surges. The deadli- est pyroclastic surge in modern times occurred in Mexico on 4 April 1982.
El Chichón Volcano sits in a remote part of Chiapas, the southernmost state in Mexico ( figure 7.28 a ). The volcano had been dormant for at least 550 years and was not consid- ered an imminent hazard. March 1982 was a month of numerous earthquakes leading up to 29 March, when an unexpected six-hour-long Plinian eruption blasted 1.4 km 3
of rock and magma into the atmosphere. The volcano had changed ( figure 7.28 b ). The eruption was surprising and the
gas-ash clouds expanded in width and overwhelmed St. Pierre (see figure 7.25 ).
How was the town of St. Pierre destroyed? The pyroclastic flow moved with hurricane speeds of about 190 km/hr (115 mph), but it was much denser than a hurricane because of its contained ash. The flow lifted roofs, knocked down most walls perpendicular to its path, twisted metal bars, and wrapped sheets of metal roofing around the scorched trunks of trees. Within the space of a couple of minutes, St. Pierre turned from a verdant tropical city to burned-out ruins covered by a foot of grey ash. Muddy ash also plastered any walls and tree trunks that were still standing.
What killed the people? Death was quick and came from one of three causes: (1) physical impact, (2) inhaling superhot gases, or (3) burns. The refugee-swollen population of St. Pierre was more than 30,000; only two people are known to have survived. One was Auguste Ciparis, a 25-year-old murderer locked in a stone-hut jail without windows and with only a small barred grat-
ing in his door. When hot gases entered his cell, he fell to the floor, suffering severe burns on his back and legs. Four days later, he was rescued; he then spent the rest of his life showing his scarred body at circus sideshows as “the prisoner of St. Pierre.” The other survivor was a man inside the same house where his family members died.
Was it safe to be on a boat in the harbor? No. The fiery hot cloud did not stop when it hit the water. Of 18 boats in the harbor, only the British steamship Roddam survived, though it was badly burned and two-thirds of its crew were dead.
Pyroclastic flows continued rolling out of Mont Pelée. St. Pierre was overwhelmed again on 20 May, but it no longer mattered. On 30 August, a pyroclastic flow moved toward the southeast and scorched Morne Rouge and four other towns, killing another 2,000 people. Despite these tragic events, at present the area is fully settled once again.
A Classic Disaster (Continued)
abb22878_ch07_174-202.indd 190abb22878_ch07_174-202.indd 190 14/12/12 3:43 PM14/12/12 3:43 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanic Processess and Killer Events 191
LAHARS Lahars are volcanic mudflows and volcanic debris flows that are fluid when moving, but begin to solidify soon after stopping. These pyroclast-carrying flows can travel at speeds up to 65 km/hr (40 mph). The combination of water plus loose pyroclasts plus steep slopes plus the pull of gravity produces lahars. The word lahar comes from Indonesia and entered the scientific language after the deadly flows from the volcano Kelut in 1586 (see figure 7.23 ).
Lahars may occur as primary events during volcanic eruptions or as secondary events months or years after erup- tion. When steep slopes are covered with loose pyroclasts, it takes only the addition of water to create lahars. Water may be available as a crater lake during eruption, such as at Kelut Volcano, or it may come later from heavy rainfalls or melting of glacial ice.
Lahars Due to Heavy Rainfall The eruptions of Mount Pinatubo in the Philippines in June 1991 featured stupendous Vulcanian events (see figure 6.3). On 15 June 1991, Typhoon Yunya with its heavy rainfall passed over the erupting volcano, sending voluminous lahars downslope and through the cities below (see figure 6.31).
Lahars Due to Melting Glacial Ice Does it take a huge eruption to kill a lot of people? No. Nevado del Ruiz in Colombia rises to an elevation of 5,400 m (more than 17,700 ft).
A 19 km 2 (7 mi 2 ) area on top of the mountain is covered by an ice cap 10 to 30 m (30 to 100 ft) thick with an ice vol- ume of about 337 million m 3 .
In November 1985, continuous harmonic tremors (earth- quakes) foretold a coming eruption. On 13 November, At
9:37 p.m., a Plinian column rose several miles high. Hot pyroclastic debris began settling onto the ice cap, causing melting. By 10 p.m., condensing volcanic steam, ice melt, and pyroclastic debris combined to send lahars down the east slopes into Chinchina, destroying homes and killing 1,800 people.
But the worst was yet to come. Increasing eruption melted more ice, sending even larger lahars flowing down the canyons to the west and onto the floodplain of the Rio Magdalena ( figure 7.29 ). At 11 p.m., the first wave of cool lahars reached the city of Armero and its 27,000 residents. These lahars had traveled 45 km (28 mi) from the mountain- top, dropping more than 5,000 m (16,400 ft) in elevation. In the steep-walled canyons, the lahars moved at rates up to 45 km/hr (28 mph), slowing as they flowed out onto the flatter land below.
A few minutes after 11 p.m., roaring noises announced the approach of successive waves of warm to hot lahars. Most of Armero, including 22,000 of its residents, ended up buried beneath lahars 8 m (26 ft) thick ( figure 7.30 ). The 22,000 unlucky people were either crushed or suffocated by the muddy lahars.
But 5,000 people did escape. How? They were higher up the slopes. A memorable video showed a man’s talking head, which appeared to be resting on top of the mudflows; the man was caught by lahars and buried to his chin as he tried to escape upslope. One step slower and he would have been completely buried and suffocated. But with a bit of digging, he was freed, shaken but unharmed.
The volcanic eruption at Nevado del Ruiz was actually rather minor. Had there not been an ice cap to melt, no harm would have been done. The November 1985 lahars were a virtual rerun of the events that occurred in that area 140 years earlier, in February 1845. The same places were buried by
Figure 7.28 El Chichón, Chiapas, Mexico. (a) Before: In September 1981, the lava dome–plugged volcano was not considered a big hazard. (b) After: During one week in 1982, the lava dome was destroyed, leaving a 1 km (0.6 mi) diameter crater. Photo (a) by René Canul D. and photo (b) by Robert I. Tilling, US Geological Survey.
(a) (b)
abb22878_ch07_174-202.indd 191abb22878_ch07_174-202.indd 191 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
192 Chapter 7 Volcano Case Histories: Killer Events
the same types of lahars. In 1845, the death toll was about 1,000, but because Colombia’s population has grown, the dead in 1985 numbered about 24,000.
Mount Rainier, Washington—On Alert Should the Seattle-Tacoma metropolitan region be concerned about lahars? Yes. Nevado del Ruiz showed that a small erup- tion on a glacier-capped volcano can be big trouble. Mount Rainier is the tallest of the Cascade Range volcanoes, at 4,393 m (14,410 ft). It stands 2,150 to 2,450 m (7,000 to 8,000 ft) above its adjacent areas and is a beautiful sentinel readily seen from throughout the Seattle-Tacoma urban
Nevado del Ruiz Volcano
Murillo
Glaciers
Lahar
Lahar
Libano Armero
Mariquita
Honda
Rio Magdalena
WE
20 km100
12 mi0 6
Figure 7.29 An eruption of Nevado del Ruiz in 1985 dropped hot pyroclastic debris onto glaciers, resulting in lahars. Source: US Geological Survey.
Figure 7.30 Most of the town of Armero, Colombia, and 22,000 of its residents lie beneath lahars up to 8 m (26 ft) thick. Photo from the US Geological Survey.
region ( figure 7.31 ). Yet Mount Rainier is number one on the danger list of many U.S. volcanologists because of its (1) great height, (2) extensive glacial cap, (3) frequent earth- quakes, and (4) active hot-water spring systems, which have weakened the mountain internally. Mount Rainier can be described as 33.6 mi 3 of structurally weak rock capped by 1 mi 3 of snow and ice; this volcanic mountain is inherently unstable. Mount Rainier is a national park and cannot be densely developed, but it nonetheless presents distinct threats to heavily populated areas. The mountain itself may fail in a massive avalanche, and/or rapidly melted ice can cause floods or lahars. Mount Rainier supports the largest glacier system of any mountain in the lower 49 states. This ice can
Figure 7.31 Mount Rainier looms on the skyline behind the Seattle-Tacoma region. Photo by Lyn Topinka, US Geological Survey.
abb22878_ch07_174-202.indd 192abb22878_ch07_174-202.indd 192 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanic Processess and Killer Events 193
be melted by magma moving up inside the mountain, even without active volcanism.
The rock record shows numerous far-reaching lahars in the last several thousand years ( figure 7.32 ). The Osceola mudflow moved about 5,600 years ago, flowing more than 120 km (75 mi) down the White River valley before spread- ing out onto the Puget Sound lowlands and into Puget Sound. It covers an area greater than 100 mi 2 to depths over 20 m (70 ft). The Osceola mudflow began as a water-saturated avalanche during summit eruptions of Mount Rainier. It transformed into a clay-rich lahar within 2 km (1.2 mi) of travel as it carried 3.8 km 3 (0.9 mi 3 ) of material at velocities up to 45 mph out across the Puget Sound lowlands. The affected area is now home to about 100,000 people. A repeat of an Osceola-size lahar could kill thousands of people. To visualize what could happen, see the 1985 lahar in Armero, Colombia (see figure 7.30 ); the Osceola event was 40 times larger than the Armero lahar.
The Electron mudflow is only 500 years old; it flowed down the Puyallup River valley for 48 km (30 mi) and also out onto the Puget Sound lowlands. Today, the region is a desirable place to live; the population is growing rapidly and building homes on top of these lahar deposits. Mount Rainier’s next major eruption may bring staggering property damage and deaths.
Warnings are possible before lahars reach towns. Moving lahars may be detected in the upper reaches of valleys near Mount Rainier by acoustic flow mon- itors (AFMs). An AFM is a seismometer that records ground vibrations at different frequencies than those generated by earthquakes or most volcanic activity. An AFM concentrates on vibrations between 10 and 300 hertz (Hz), whereas seismometers recording earthquakes and volcanoes commonly focus on waves between 0.5 and 20 Hz. When data from AFMs cross critical values, they are transmitted by radio to emergency centers, and they can also trigger auto- matic warning devices.
DEBRIS AVALANCHES A tall stratovolcano is a beautiful sight and appears to be a mountain of strength. In reality, though, many centuries of forceful intrusions of magma into stratovolcanoes riddle them with fractures, creating planes of weakness. Hot water and gases rising through fractures chemically decompose the volcanic rock over time and weaken it. The frac- tures and rotten rock can lead to massive failures: sector collapses that flow downslope as debris avalanches (see Mount Shasta photo in figure 7.18 and map in figure 7.21 ). A debris avalanche deposit is composed of huge blocks of the volcano within a matrix of finer-grained material (see sector collapse of north side of Mount St. Helens in
figure 7.11 c and debris avalanche material choking the Toutle River in figure 7.16 ).
A volcano sector collapse may be triggered by the injec- tion of fresh magma that inflates a volcano; by forceful expansion of water in contact with magma inside a volcano; or by an earthquake. Debris avalanches are responsible for 2% of volcano-caused deaths.
INDIRECTFAMINE Volcanoes affect humans not only directly, but also indirectly. Volcanism can reduce agricultural output, weaken or kill livestock, and weaken humans, setting the stage for famine.
Laki, Iceland, Fissure Eruption of 1783 During the summer of 1783, the greatest lava eruption of historic times poured forth near Laki in Iceland. After a week of earthquakes, on the morning of 8 June 1783, a 25 km (16 mi) long fissure opened, and basaltic lavas gushed for 50 days at 5,000 m 3 /sec. To better appreciate this volume of magma, consider that North America’s mightiest river, the Mississippi, empties into the Gulf of Mexico at about three times this volume. When the eruption ended, an area of 565 km 2 (218 mi 2 ) was buried beneath 13 km 3 (3 mi 3 ) of basaltic lavas. The volume of ash and larger airborne frag- ments totaled another 0.3 km 3 .
0
0 6 mi
5 10 km
122°
Mount Rainier National Park
Longmire Ashford
Elba
Glaciers
Sumner
Puget Sound
Buckley
Orting
Puyallup
Enumclaw
Tacoma
Auburn
Nisqually River
Carbon Rive r
White River
West Fork
5,000 years old
Electron mudflow
Osceola mudflow
500 years old
M ou
nt ai
n
F
ro nt
Puyallup River
Pu ge
t S ou
nd Lo
wla nd
Figure 7.32 Map showing the area covered by two of the many lahars that have flowed from Mount Rainier. Source: D. R. Crandell and D. R. Mullineaux, “Volcanic Hazards at Mt. Rainier, Washington,” 1967, in US Geological Survey Bulletin 1238.
abb22878_ch07_174-202.indd 193abb22878_ch07_174-202.indd 193 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
194 Chapter 7 Volcano Case Histories: Killer Events
The 50 days of eruption were accompanied by the release of an enormous volume of gases that enshrouded Iceland and much of northern Europe in a “dry fog” or blue haze. This haze was rich in SO 2 (one of the visible components of today’s urban smog) and an unusually large amount of fluorine. The gases slowed the growth of grasses and increased their fluorine content. An Icelandic farmer named Jon Steingrimmson wrote:
The hairy sand-fall and sulfurous rain caused such unwhole- someness in the air and in the earth that the grass became yellow and pink and withered down to the roots. The animals that wandered around the fields got yellow-colored feet with open wounds, and yellow dots were seen on the skin of the newly shorn sheep, which had died.
The volcanic gases helped kill 75% of Iceland’s horses and sheep and 50% of the cattle. The resulting famine weakened the Icelandic people, and about 20% of the population (10,000 people) died. In today’s world of instant communication and rapid air transport, these deaths would have been avoided.
Tambora, Indonesia, 1815 The most violent and explosive eruption of the last 200 years was another Indonesian event; it came from Tambora Volcano on Sumbawa Island in April 1815. After three years of moderate activity, on 5 April, a Plinian eruption column shot up 33 km (20 mi) and carried out 12 km 3 (2.9 mi 3 ) of pumice in just two hours. On 10 April, an even more powerful Plinian eruption blasted up to 44 km (27 mi) high for three hours. The magma exited with so much force that it eroded and widened the vent in the volcano, thus cutting off the focused energy that drove the Plinian column. The eruption column stopped, and the widened vent lay open; with its insides exposed, the volcano now spilled its guts. On 11 April, about 50 km 3 (12 mi 3 ) of magma poured out of the caldera in over- whelming pyroclastic flows. The week-long eruption saw about 150 km 3 (36 mi 3 ) of magma burst forth. Tambora once stood 4,000 m (13,000 ft) high, but now its elevation was reduced to 2,650 m (8,700 ft) with a 6 km (3.7 mi) wide caldera that was over 1 km (0.6 mi) deep. The volcanic explosions were audible 2,600 km (1,600 mi) away, and volcanic ash fell 1,300 km (800 mi) from Tambora. On Sumbawa Island, pyroclastic flows killed at least 10,000 people. They also destroyed the feudal kingdoms of Sanggar and Tambora, leading to the erasure of the Tambora language, the easternmost Austro-Asiatic language.
The eruption of Tambora was responsible for an esti- mated 117,000 deaths: about 10% killed by the eruption and 90% dying slowly at the end of a chain reaction. Pyroclastic fallout devastated crops, which led to famine and weakened people, making them more susceptible to disease, and then the diseases killed them. But this was not just another Indonesian disaster. The Plinian eruptions of April 1815 so affected global climate that 1816 is known as “the year with- out a summer.” The climatic effects of the eruption are discussed in chapter 12.
GAS It is not just gas-powered magma that kills; gas can be deadly all by itself. Gases are a continuous product of volcanism but even nonerupting volcanoes can release significant volumes of gas.
Killer Lakes of Cameroon, Africa Spreading centers commonly begin as three-armed rifts meeting at a triple junction (see figure 4.7). In northeast Africa, two rift arms have spread apart enough to create the Red Sea and the Gulf of Aden, while the third arm has failed, so far, to open the East African Rift Valley into another new ocean basin (see figure 4.5). Failed rifts that do not open up enough to become spreading centers are common (e.g., figure 7.35 ). If a rift fails to open a new ocean basin, must it stop all activity? No.
Cameroon sits near the equator in western Africa. It hosts a string of crater lakes running in a northeasterly trend. Prolific rainfalls fill the lakes and combine with the hot tem- peratures to cover the countryside with greenery. Lake Nyos is one of these crater lakes, filled with beautiful, deep-blue water. This topographically high crater is only several hundred years old. It was blasted into the country rock by explosions of volcanic gases and is 1,925 m (6,310 ft) across at its greatest width and as deep as 208 m (680 ft).
At about 9:30 p.m. on 21 August 1986, a loud noise rumbled through the Lake Nyos region as a gigantic volume of gas belched forth from the crater lake and swept down the adjacent valleys ( figure 7.36 ). The dense, “smoky” rivers of gas were as much as 50 m (165 ft) thick and moving at rates up to 45 mph. The ground-hugging cloud swept outward for 25 km (16 mi). Residents of four villages overwhelmed by the gaseous cloud felt fatigue, light-headedness, warmth, and confusion before losing consciousness. After 6 to 36 hours, about half a dozen people awoke from their comas to find themselves in the midst of death: 1,700 asphyxiated people; 3,000 dead cattle; and not a bird or insect alive, nor any other animal. Yet the luxuriant plants of the region were unaffected.
This shocking event raised numerous questions. What was the death-dealing gas? What was the origin of the gas? How did the gas accumulate into such an immense volume? What triggered the gas avalanche? Is this event likely to hap- pen again?
What was the death-dealing gas? After a lot of effort to identify some exotic lethal gas or toxic substance as the cause of the tragedy, the killer gas turned out to be simply carbon dioxide. This is the same gas we drink in sparkling spring water, soda pop, and champagne. Its toxicity at Nyos is explained by the principle set forth in 1529 by the German physician Theophrastus von Hohenheim (Paracelsus). The principle of Paracelsus states: the dose alone determines the poison. A gas does not have to be poisonous, just abundant. Life in the Nyos region was subjected to the same conditions we recreate inside the fire-extinguisher cylinders in our buildings. Fire extinguishers are loaded with carbon dioxide,
abb22878_ch07_174-202.indd 194abb22878_ch07_174-202.indd 194 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcanic Processess and Killer Events 195
Side Note Death at Ashfall, Nebraska Ten million years ago, the area around Ashfall, Nebraska, held water holes within a savanna setting, a warm, flat grassland similar to some classic wildlife areas in Africa today. Large herds of animals migrated to the water holes to drink: three-horned deer, giant camels, three-toed horses, oreodonts, four-tusked elephants, weasels, bear dogs, rhinoceroses, and many more species. Their daily routines changed for the worse one day when a huge volcanic eruption blasted forth 1,300 km (800 mi) away. The eruption came from the Yellowstone hot spot, but 10 million years ago, it sat beneath Idaho (see figure 6.38). Winds carried volcanic ash from Idaho and blan- keted Nebraska with a layer of ash about 1 ft thick. After its initial deposition, local winds picked up and blew ash around in gray blizzards. Large amounts of reworked ash settled in the water holes.
What effects can cool, loose, ultrafine volcanic ash have on life? All the animals inhaled volcanic ash for days, weeks, and months, causing health problems. The high magnification of a scanning electron microscope reveals that volcanic ash is composed of sharp, jagged, angular pieces of glass and rock ( figure 7.33 ), which are irritants inside living bodies. Breathing becomes difficult, and respiratory problems develop. Fossil bones of large animals at Ashfall show irregular growth, evidence that they were not getting enough oxygen to grow normal bones.
Fossil preservation at Ashfall is superb, with whole animal skele- tons still joined together as they were in life. The layers of fossil-con- taining ash show the death sequence. In the lowest ash layer are the remains of the first to die—birds and turtles. In overlying ash layers are the fossils of musk deer and small carnivores. Some of the next animals
to perish were the horses and camels. A herd of about 100 rhinocer- oses kept returning to the water holes, kicking up and breathing vol- canic ash clouds each time, until they too died ( figure 7.34 ). Their fossils include a mother rhino who died before her suckling youngster lying next to her. As ash continued to blow about, it ultimately buried the water hole death sites. You can see the herds of animals, partially excavated and available for viewing, at Ashfall Fossil Beds Historical Park in Antelope County, Nebraska.
Figure 7.34 Skeletons of rhinoceroses killed 10 million years ago by breathing fine volcanic ash for days, Ashfall, Nebraska. Photo by Pat Abbott.
Figure 7.33 Volcanic ash. (a) Glass fragment from Katmai, Alaska (scale = 0.04 mm). (b) Tail of cooled droplet from Kilauea, Hawaii (scale = 0.01 mm). (c) Angular glass fragment, Mount Mayon, Philippines (scale = 0.01 mm). Source: SEM photos from Grant Heiken, Atlas of Volcanic Ash, Smithsonian Press.
which does not put out flames directly; because CO 2 is heavier than air, it deprives fire of oxygen, thus causing flames to die out. Animal life in the Nyos area was extin- guished in the same fashion.
What was the origin of the gas? It had a volcanic origin, leaking upward from underlying basaltic magma. A 1,600 km (1,000 mi) long string of volcanoes, the Cameroon volcanic line, trends northeastward through several Atlantic Ocean
abb22878_ch07_174-202.indd 195abb22878_ch07_174-202.indd 195 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
196 Chapter 7 Volcano Case Histories: Killer Events
islands and then on land through northeastern Nigeria and northwestern Cameroon. Interestingly, this is the location of the triple junction of spreading centers that ripped apart this section of Gondwanaland, helping give the distinct outlines to the Atlantic margins of South America and Africa ( figure 7.35 ). The two successful spreading arms are still widening the South Atlantic Ocean. The failed rift is occu- pied by the line of volcanism that includes the crater that forms Lake Nyos; it is not a volcanic mountain but a crater blasted through bedrock by largely gaseous explosions. The volcanic activity is not seafloor spreading per se; rather, it is a “wannabe” ocean basin that never made it but has not given up totally.
How did the gas accumulate into such an immense volume? Lakes by their nature are stratified bodies of water. Their water layers differ in density, one stacked on top of another. (This is a smaller-scale example of the density dif- ferentiation discussed for the whole Earth in chapter 2.) Carbon dioxide, given off by basaltic magma at depth, rises into the bottom waters of Lake Nyos, is dissolved into the heavier, lower water layer, and is held there under the pres- sure of the overlying water ( figure 7.36 ). As the amount of CO 2 in the lake-bottom water increases, it becomes more unstable. When CO 2 bubbles form, they rise with increas- ing speed, setting off a positive feedback chain of events leading to more and more bubble formation and rise. Volca- nologist Youxue Zhang calculated that the gas eruption was moving about 200 mph when it reached the lake surface.
The event of 21 August 1986 released about 0.15 km 3 of gas in about one hour. It was like a large-scale erupting cham- pagne bottle, where removal of the cork causes a decrease in pressure, allowing CO 2 to escape in a gushing stream. About 66% of the dissolved gases escaped. After the event, the lake level was 1 m lower, and the water was brown from mud and dead vegetation stirred up from the bottom ( figure 7.37 ).
What triggered the gas avalanche? Many suggestions have been made, including volcanic eruption, landslide, earthquake, wind disturbance, or change in water tempera- ture with resultant overturn of lake-water layers. It is inter- esting to note that a similar event occurred two years earlier at Lake Monoun on 15 August 1984. This was a smaller event, but it killed 37 people. Both events were in August, the time of minimum stability in Cameroon lake waters. Is this a coincidence, or is this a normal overturning of lake water during the rainy season?
SOUTH AMERICA
AFRICA
Lake Monoun
Lake Nyos
Figure 7.35 Schematic map of Africa and South America splitting apart 135 million years ago. Note the failed rift extending into Africa (upper right corner).
Volcanic gas
Lake Nyos
“River” of CO2 about 50 m thick
Figure 7.36 Schematic cross-section of Lake Nyos. Vol- canic gas is absorbed by the deep-water layer. In 1986, when bottom water was disturbed, 0.15 km 3 of CO 2 gas erupted out of the lake and poured down river valleys for an hour or more in a 50 m thick cloud. Virtually all animal life was killed; plants were unaffected. Solid lines show gas flow; dashed lines are water drops. Diagram after Y. Zhang, 1996, Nature, 379, 57–59.
Figure 7.37 The water of Lake Nyos was still muddy 10 days after a huge volume of gas escaped in 1986. Photo from US Geological Survey .
abb22878_ch07_174-202.indd 196abb22878_ch07_174-202.indd 196 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
VEIs of Some Killer Eruptions 197
Is this event likely to happen again? Definitely. The Lake Nyos gas escape left behind 33% of the CO 2 , and more is constantly being fed through the lake bottom. In about 20 years, the lake water could again be oversaturated with CO 2 . The same loss of life will occur again unless remedial actions are taken. Degassing pipes have been installed to allow high-pressure gas to shoot out of the lake as a fountain of gassy water. This could prevent the CO 2 concentrations from building up to explosive levels.
As this situation has become better known, other similar lakes have been recognized. For example, the giant Lake Kivu that straddles the border between Rwanda and Congo holds more than 350 times as much gas as Lake Nyos.
LAVA FLOWS Lava flows are common and impressive, but they are respon- sible for less than 1% of volcano-caused deaths. Why don’t lava flows kill more people? Usually they move too slowly— but not always.
Nyiragongo, Zaire, 2002 As East Africa slowly rifts away from the African continent (see figure 4.5), magma rises to build stratovolcanoes such as Mount Nyiragongo in the East African Rift Valley. Nyiragongo has a long-lived lava lake in its summit crater. On 17 January 2002, lava flowed rapidly down the slopes of the volcano, killing more than 100 people living on the mountain. Upon reaching flatter ground, the lava flows slowed but moved relentlessly toward Lake Kivu. The city of Goma lay in the path of the oncoming lava: 500,000 residents plus uncounted thousands of civil war refugees from Rwanda lived there. Lava reached the lake, but it first flowed through the heart of Goma, destroying about 25% of the buildings and forcing the war refugees to flee again.
How were the lava flows able to catch and kill so many people? The lava had unusually low viscosity. In 1977, Nyiragongo lava flows had exceptionally low SiO 2 content, about 42%. (Compare this value to table 6.5.) The low-viscosity lava in 1977 flowed down the volcano slopes at about 60 km/hr (40 mph), killing an estimated 300 people.
An additional concern is the tremendous volume of car- bon dioxide and methane gas held in the deep water of Lake Kivu. A large disruption of the bottom waters, as by an enter- ing lava flow or eruption on the lake bottom, could cause a gas release affecting the 2 million people living along the shore of Lake Kivu.
VEIs of Some Killer Eruptions Does the total energy involved in a volcanic eruption corre- late well with number of deaths? Not necessarily. The volca- nic explosivity index (VEI), in chapter 6, is a semiquantitative
approach to estimating the magnitude of explosive eruptions based on volume of material erupted and eruption-column height. Table 7.2 lists VEIs for some of the deadly events we have examined. Note that some of these events had low VEIs; they killed with a relatively small-volume pyroclastic flow, melted glacier ice, and gas escape without magma.
How frequent are eruptions at specific VEI magnitudes? Somewhere on Earth, a VEI 2 event occurs every few weeks, a VEI 3 happens several times a year, a VEI 4 erupts once or twice a year, a VEI 5 happens about once per decade, and a VEI 6 blasts forth once or twice a century. The bigger the eruptions, the less frequently they occur. As the human pop- ulation continues its rapid growth, increasing numbers of people move into volcano hazard zones. The need for accu- rate prediction of eruptions is becoming ever more pressing.
How much harm can a volcano do? The volcano Toba may have driven the human race almost to extinction about 74,000 years ago. The eruption was a resurgent caldera event that ejected 2,800 km 3 (670 mi 3 ) of rock, magma, and ash in a mega-eruption with a VEI of 8 ( table 7.2 ). The size of this eruption is equivalent to a combined 560 of the Mount Pinatubo eruptions in 1991. The scar from the Toba eruption is the 100 km (60 mi) long Lake Toba lying near the equator at 2.5° North latitude on Sumatra. Low-latitude eruptions are the most dangerous because they spread debris around more of the world, as gases and fine ash choke the atmosphere and affect life globally.
DNA studies of humans alive today suggest we are descended from a small population of 1,000 to 10,000 people. The average rates of genetic mutation within our own DNA suggest that the time when human population was severely reduced was about the same time as the eruption of Toba. The regional devastation and global cooling of the climate caused by the Toba eruption may have made life so difficult that humans were almost forced into extinction.
TABLE 7.2 VEIs of Notable Volcanic Disasters (Volcanic Explosivity Index)
VEI Volcano 8 Yellowstone, 600,000 years ago;
Toba, 74,000 years ago
7 Tambora, 1815
6 Vesuvius, 79; Krakatau, 1883; Pinatubo, 1991
5 St. Helens, 1980
4 Pelée, 1902
3 Nevado del Ruiz, 1985
2 —
1 —
0 Lake Nyos, 1986
abb22878_ch07_174-202.indd 197abb22878_ch07_174-202.indd 197 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
198 Chapter 7 Volcano Case Histories: Killer Events
Volcano Monitoring and Warning Can we monitor the activity of a volcano and provide advance warning before a large eruption? Efforts to do so have met with both failure and success.
LONG VALLEY, CALIFORNIA, 1982 In the Long Valley–Mammoth Lakes area of California, abundant melting of crustal rock occurs, although no classic hot spot exists there. About 760,000 years ago, a colossal eruption blew out more than 600 km 3 (150 mi 3 ) of magma, generating pyroclastic flows that covered an east-central California area greater than 1,500 km 2 (580 mi 2 ) with pyro- clastic debris (called Bishop Tuff) up to hundreds of meters thick. Immediately after the magma blasted out, Earth’s sur- face dropped nearly 2 km (more than 1 mi) into the void to form the Long Valley caldera ( figures 7.38 and 7.39 ). One pyroclastic lobe flowed 65 km (40 mi) down the Owens Val- ley. Before the eruption, the magma body is estimated to have had a diameter of 19 km (12 mi), with its roof 5 km (3 mi) below the surface.
0
0
10 mi
16 km
* *
*
*
*
*
Long Valley caldera
Mammoth Mountain
Inyo craters
Mono craters
Owens Valley
Bishop
Mono Lake
Mammoth Lakes
395
Bishop Tuff (arrows indicate directions of flow)
Volcanic Rocks
Earthquakes > 6 magnitude (1978-1995)
Mono-Inyo craters volcanic chain
395
Figure 7.38 Map showing the Long Valley caldera formed by massive eruptions. The brown areas with red arrows are composed of Bishop Tuff, uneroded remains of pyroclastic debris from the last major eruption. The section of Highway 395 shown here lies just north of the section pictured in figure 5.20.
Huge eruptions are rare, but these giant continental calderas have fairly frequent small eruptions. There were erup- tions in Long Valley 600 years ago and in Mono Lake just 150 to 250 years ago. Today, the main magma body is about 10 km (6 mi) in diameter and around 8 km (5 mi) deep ( figure 7.40 ).
On 25–26 May 1980, one week after the catastrophic eruption of Mount St. Helens, Long Valley was shaken by numerous earthquakes within 48 hours: four of magnitude 6, dozens of magnitude 4 to 5, and hundreds of smaller seisms. In the resort town of Mammoth Lakes, foundations and walls cracked, chimneys fell, and pantry and store shelves dumped their goods. Monitoring by the U.S. Geological Survey showed the resurgent dome had risen 25 cm (10 in) in late 1979–early 1980 (see dome in figure 6.43). The dome rose another several inches by early 1982, accompanied by swarms of earthquakes. Some magma that was 8 km (5 mi) deep in 1980 had risen to within 3 km (2 mi) of the surface by 1982.
Was a volcanic eruption imminent? What should be done? The affluent town of Mammoth Lakes draws most of its income from tourism; it has a year-round population of 5,500 but adds another 20,000 during winter ski season. Would issuing a formal warning of volcanic hazard do good or harm? On 27 May 1982, the U.S. Geological Survey issued a Notice of Potential Volcanic Hazard, the lowest level of alert. House prices fell 40% overnight and tourist visits dropped dramatically. Home and business owners erupted, but the volcano did not.
In the early 1990s, trees on Mammoth Mountain began dying as large amounts of carbon dioxide (CO 2 ) rose up from the underlying magma and killed them. At the same time, small earthquakes resumed and the ground surface began rising. These phenomena can occur for decades or centuries; however, at large calderas, they do not always mean an erup- tion is imminent. Many residents remain angry about the “false alarm” of 1982, while many volcanologists and emer- gency planners are hesitant to issue another volcano warning. Residents are advised to follow the motto: prepare for the worst, but hope for the best.
To get a good view of the giant caldera that is Long Valley, look over your left shoulder as you ride up the chairlifts at the Mammoth Mountain ski resort. The big, dry valley below is the caldera ( figure 7.39 ).
MOUNT PINATUBO, PHILIPPINES, 1991 A volcano-warning success story occurred in the Philippines in 1991 before the climactic eruption of Mount Pinatubo on 15 June. The volcanic eruption was the largest in the 20th century to occur near a heavily populated area. Nearly 1 million people, including 20,000 U.S. military personnel and their dependents, lived in the danger zone.
In March 1991, Mount Pinatubo awoke from a 500-year- long slumber as magma moved upward from a depth of 32 km (20 mi), causing thousands of small earthquakes,
abb22878_ch07_174-202.indd 198abb22878_ch07_174-202.indd 198 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Volcano Monitoring and Warning 199
Figure 7.39 The large valley in the center and center-right is the caldera complex of Long Valley, California. Photo by John S. Shelton.
00 1 2 3 4 5
10
15
20
1
2
3
4
5
6
7
8
9
10
1980–1982 intrusion
SW NE
Sediments
Bishop Tuff
1980 magma rise
D ep
th (
km )
D ep
th (
m i)
Solidified part of magma chamber
Magma chamber
Long Valley caldera
Area of 1980 uplift
Figure 7.40 This cross-section oriented northeast–southwest through the Long Valley caldera shows the size and depth of the magma body. A tonguelike intrusion moved up the southern edge of the caldera in 1980–1982.
abb22878_ch07_174-202.indd 199abb22878_ch07_174-202.indd 199 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
200 Chapter 7 Volcano Case Histories: Killer Events
creating three small steam-blast craters, and emitting thou- sands of tons of sulfur dioxide–rich gas. U.S. and Philippine volcanologists and seismologists began an intense monitoring program to anticipate the size and date of a major eruption. On 7 June, magma reached the surface but had lost most of its gas (like a stale glass of soda pop), so the magma simply oozed out to form a lava dome (see figure 7.14 ). Then, on 12 June (Philippine Independence Day), millions of cubic meters of gas-charged magma reached the surface, causing large explosive eruptions. It was time to get out of the volcano’s killing range! The message to speed up the evacuation spread quickly and loudly. Virtually every person, and every movable thing, left hurriedly. On 15 June, the cata- clysmic eruption began ( figure 7.41 ). It blew more than 5 km 3 (1 mi 3 ) of magma and rock up to 35 km (22 mi) into the atmosphere, forming an ash cloud that grew to more than 480 km (300 mi) across. The airborne ash blocked incoming sunlight and turned day into night. Pyroclastic flows of hot ash, pumice, and gas rolled down the volcano flanks (see figure 6.3) and filled valleys up to 200 m (660 ft) deep. Then, as luck would have it, a typhoon (hurricane) arrived and washed tremendous volumes of volcanic debris downslope as lahars (see figure 6.31).
How successful was the advance warning? Although almost 300 people died, it is estimated that up to 20,000 might have died without the forceful warnings. The score- card for the monitoring program from March to June 1991 shows that a monitoring expense of about $1.5 million saved 20,000 lives and $500 million in evacuated property, including airplanes. What a dramatic and cost-effective success!
SIGNS OF IMPENDING ERUPTION Several phenomena are being evaluated as signs of impend- ing eruption. We need to determine if they are reliable enough to justify evacuating people out of a volcanic-hazard zone. Phenomena being studied include seismic waves, ground deformation, and gas emissions.
Seismic Waves As magma rises up toward the surface, it causes rocks to snap and break, thus sending off short-period seismic waves (SP) with typical periods of 0.02 to 0.06 second. Magma on the move through an opened conduit generates longer-period seismic waves (LP) with periods of 0.2 to 2 seconds. In 1991, during the two weeks before Mount Pinatubo erupted, there were about 400 LP events a day coming from about 10 km (6 mi) deep. Apparently, the LP events were recording the arrival of new magma moving in and loading the volcano for eruption.
Further study shows that a volcano in action causes a variety of earthquakes that generate seismic waves with dif- ferent periods. These seismic waves (acoustic emissions) produce a record similar to a symphony orchestra. Recogni- tion of different seismic waves could develop into a way of forecasting eruptions. Volcanic processes include (1) cre- ation and propagation of fractures in rock, (2) active injec- tion and movement of magma, (3) degassing, and (4) changes in pore-fluid pressures. Seismic waves from volcanic activity include (1) high-frequency waves, (2) low-frequency waves, (3) very-low-frequency waves, (4) tremors of continuous low-frequency vibrations, and (5) hybrid mixtures. The goal is to relate each seismic-wave type to a specific volcanic activity. In effect, the work is like distinguishing the sound of the flute or the clarinet from the many sounds produced by a symphony orchestra.
Ground Deformation The ground surface rises up and sinks down as magma moves up or withdraws. Ground deformation can be measured by tilt meters or strain meters placed in the ground and by elec- tronic distance meters. In recent years, satellites have been using radar to measure movements of the ground over time. For example, between 1996 and 2000, a 15 km (9 mi) wide area on the flanks of the Three Sisters volcanoes in Oregon bulged upward 10 cm (4 in) as about 21 million m 3 of magma rose to within 6 to 7 km (3.7 to 4.3 mi) below the ground surface. Now, global positioning system (GPS) stations have been set in the area to add more data about ground deforma- tion. The more the ground rises, the more likely it is that some magma will break through to the surface and erupt.
Gas Measurements As magma rises toward the surface, the pressure on it drops and dissolved gases escape. For example, at Mammoth Mountain next to the Long Valley caldera in California (see figure 7.38 ), CO 2 is escaping from the magma. A decade ago, more than 1,000 tons a day were oozing through the surface, killing trees and causing worry about an impending eruption. Now, CO 2 releases have declined to about 300 tons a day, suggesting that an impending eruption is less likely. However, this interpretation could be misleading. In 1993, at Galeras Volcano in Colombia, a decrease in gas emissions was interpreted as meaning an eruption was less likely. But, in fact, it meant that the volcano had become plugged by its
Figure 7.41 The 15 June 1991 Plinian-type eruption of Mount Pinatubo lasted 15 hours, sending pyroclastic flows downslope. VEI = 6. Photo by Robert Lapointe, US Air Force.
abb22878_ch07_174-202.indd 200abb22878_ch07_174-202.indd 200 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
Questions for Review 201
sticky magma, and gas pressure was building toward the eruption that killed seven volcanologists. So, either an increase or a decrease in gas emissions can be bad. More research must be done.
VOLCANO OBSERVATORIES As volcanoes continue to burst out with damaging and killing eruptions, many countries are responding by establishing and staffing volcano observatories to provide warnings before big eruptions. In the 20th century, the United States
experienced powerful eruptions in four states: Alaska, California, Hawaii, and Washington. There are at least 65 active or potentially active volcanoes in the United States. Watching these volcanoes for signs of activity has led the U.S. Geological Survey to establish a Volcano Hazards Program that includes five volcano observatories: Alaska (AVO), California (Cal Vo), Cascades (CVO), Hawaiian (HVO), and Yellowstone (YVO). Each of these observatories maintains its own website to report current activity. Their observations are also presented using an alert system ( table 7.3 ).
TABLE 7.3 Volcanic-Alert Levels, US Geological Survey
Normal Typical background activity of a volcano in a noneruptive state
Advisory Elevated unrest above known background activity
Watch Heightened/escalating unrest with increased potential for eruptive activity or minor eruption underway
Warning Highly hazardous eruption underway or imminent
Summary Spreading centers provide such ideal settings for volcanism that 80% of all extruded magma occurs there. Spreading centers sit on top of the asthenosphere, which yields basaltic magma that rises to fill fractures between diverging plates. Basaltic volca- noes may be successfully colonized both at spreading centers (e.g., Iceland) and at oceanic hot spots (e.g., Hawaii).
Subduction-zone eruptions involve basaltic magma con- taminated by crustal rock to yield water-rich, highly viscous magma containing trapped gases. Their explosive eruptions make the news (e.g., St. Helens, Unzen, and Pinatubo) and the history books (e.g., Santorini, Vesuvius, and Krakatau). Transform faults and continent-continent collisions have lit- tle or no volcanism associated with them.
The historic record tells of about 275,000 deaths by volcano in the last 500 years. The deadliest processes have been pyroclastic flows, tsunami, lahars, and indirect effects leading to famine. Gas-powered pyroclastic flows can move at speeds up to 150 mph with temperatures over 1,300°F and for distances over 30 mi; examples are Mont Pelée and El Chichón. Pyroclastic debris and water combine and flow downslope as lahars at speeds up to 30 mph and for distances up to 45 mi, killing thousands at Nevado del Ruiz and pre- senting a hazard to Seattle-Tacoma from Mount Rainier. Volcano-generated tsunami were mega-killers at Krakatau. Sectors of volcanic cones can collapse, producing giant debris avalanches that bury entire landscapes up to 30 mi away (e.g., Mount Shasta). Giant eruptions from continental calderas can erupt more than 1,000 times as much magma as a typical volcano (e.g., Long Valley).
A volcano may be active for millions of years, but centu- ries may pass between individual eruptions. The timescale of an active volcano must be considered by people living nearby.
It is possible to monitor a volcano and give advance warning of a major eruption. At Mount Pinatubo in the Phil- ippines, early warning saved up to 20,000 lives.
Terms to Remember failed rift 194 fissure 193 n uée ardente 188
pyroclastic fall 186 pyroclastic surge 186 sector collapse 193
Questions for Review 1. How many years might one subduction zone operate? One
volcano? How many years might pass between eruptions at an active volcano?
2. Explain why it is relatively safe to watch the eruption of a Hawaiian volcano but dangerous to watch a Cascade Range volcano.
3. What is sector collapse? What is a debris avalanche? 4. Draw a plate-tectonic map and explain the origin of the
Cascade Range volcanoes. 5. Draw a series of cross-sections and explain the sequence of
events in the Mount St. Helens eruption in 1980. 6. What volcanic processes have killed the most people in the
last 500 years? 7. Name four ways of creating pyroclastic flows.
abb22878_ch07_174-202.indd 201abb22878_ch07_174-202.indd 201 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
202 Chapter 7 Volcano Case Histories: Killer Events
18. What signs of impending eruption are produced by an active volcano?
19. What is column collapse? How does it generate pyroclastic flows?
20. What health hazards are associated with ash fall? What is the composition of most volcanic ash?
Questions for Further Thought 1. Is a Cascade Range volcano likely to have a major eruption
during your lifetime? 2. Is it wise for towns near Mount Shasta to keep growing?
What should be done about this situation? 3. Is it wise to build in river valleys below Mount Rainier, even
tens of miles away?
8. Why do pyroclastic flows travel so fast? How do they kill? 9. Is a Plinian eruption most dangerous when it is the strongest? 10. Can pyroclastic flows travel outward in all directions
simultaneously? (See pyroclastic surges). 11. How far can a pyroclastic flow travel over water and still be
hot enough to kill people? 12. Draw a cross-section and explain how lahars form and move.
How do they kill? 13. What four factors combine to produce lahars? 14. Explain the hazard that Mount Rainier presents to the Seattle-
Tacoma region. 15. Draw a cross-section and explain the sequence of events at an
African killer lake, such as Nyos. 16. How can an eruption with a low VEI (low magnitude
explosivity) rating kill thousands of people? 17. How do the ages vary along a line of subduction-zone
volcanoes compared to a line of hot-spot volcanoes?
abb22878_ch07_174-202.indd 202abb22878_ch07_174-202.indd 202 14/12/12 3:44 PM14/12/12 3:44 PM
G A R R E T T , M E G A N 1 3 2 4 T S
- 9780078022876_ch04_079-109_Print
- 9780078022876_ch05_110-142_Print
- 9780078022876_ch06_143-173_Print
- 9780078022876_ch07_174-202_Print
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile () /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.4 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket true /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize false /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo true /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile (None) /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages false /ColorImageDownsampleType /Average /ColorImageResolution 300 /ColorImageDepth 8 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /FlateEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages false /GrayImageDownsampleType /Average /GrayImageResolution 300 /GrayImageDepth 8 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /FlateEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages false /MonoImageDownsampleType /Average /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile () /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.4 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket true /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize false /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo true /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile (None) /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages false /ColorImageDownsampleType /Average /ColorImageResolution 300 /ColorImageDepth 8 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /FlateEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages false /GrayImageDownsampleType /Average /GrayImageResolution 300 /GrayImageDepth 8 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /FlateEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages false /MonoImageDownsampleType /Average /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile () /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.4 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket true /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize false /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo true /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile (None) /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages false /ColorImageDownsampleType /Average /ColorImageResolution 300 /ColorImageDepth 8 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /FlateEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages false /GrayImageDownsampleType /Average /GrayImageResolution 300 /GrayImageDepth 8 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /FlateEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages false /MonoImageDownsampleType /Average /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice
<< /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile () /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Error /CompatibilityLevel 1.4 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket true /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails true /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 524288 /LockDistillerParams true /MaxSubsetPct 100 /Optimize false /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo true /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Remove /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile (None) /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages false /ColorImageDownsampleType /Average /ColorImageResolution 300 /ColorImageDepth 8 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /FlateEncode /AutoFilterColorImages false /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages false /GrayImageDownsampleType /Average /GrayImageResolution 300 /GrayImageDepth 8 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /FlateEncode /AutoFilterGrayImages false /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages false /MonoImageDownsampleType /Average /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /CreateJDFFile false >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice