Thus far we focused on understanding the characteristics and origins of tectonic structures (such as faults, folds, and fabrics), on relations among deformation, stress and strain, and associated processes. Geologists use the term tectonics, derived from the Greek tektos meaning builder, to refer to the physical processes that yield regional to global scale geologic features. Studies in tectonics consider such issues as the origin and evolution of mountain belts (Figure 11.1), the growth and evolution of continents, the formation and destruction of ocean floor, the development of sedimentary basins, and, particularly relevant to today's human society, the causes of earthquakes and volcanoes. Over the next several chapters, we examine this “big-picture”, and how it connects geologic structures on all scales. Starting on this large topic may also benefit from a look at this short narrated video (by IRIS) on plates and tectonics.

Figure 11.1. The Himalayas as seen from space. This spectacular range separates the Indian subcontinent with major south-flowing rivers (left) from Asia’s high-elevation Tibetan Plateau with many lakes (right). [NASA]

We begin by examining whole-Earth structure, meaning the internal layering of our planet. Geologists subdivide the Earth’s insides into layers in two different ways, as we learned earlier. First, based on studying the velocity of seismic waves that pass through the Earth, we subdivide the entire planet into the crust, the mantle, and the core, named in sequence from surface to center. Second, based on rock rheology (the response of rock to stress), we subdivide the outer several hundred kilometers of the Earth into the lithosphere and asthenosphere. The lithosphere is the outermost of these rheologic layers and forms Earth’s rigid shell. The lithosphere, which consists of the crust and the outermost mantle, overlies the asthenosphere, a layer that deforms easily (meaning that, though solid, it flows like a viscous fluid). A background on whole Earth structure sets the stage for introducing plate tectonics theory. According to this unifying theory, today’s lithosphere consists of about 20 discrete pieces, or plates, which slowly move relative to one another. In our discussion, we will describe the nature of plates, the boundaries between them, the geometry of plate movement, and the forces that drive plate movement.

EARTH’S INTERNAL LAYERS

Before the 20th C, little was known of Earth’s interior except that it must be hot enough, locally, to generate volcanic lava. This lack of knowledge was exemplified by the 1864 novel Journey to the Center of the Earth, in which the French author Jules Verne speculated that the Earth’s interior contained a network of caverns and passageways through which intrepid explorers could gain access to the planet’s very center. Our picture of Earth’s insides changed in the late nineteenth century, when researchers compared the gravitational pull of a mountain to the gravitational pull of the whole Earth and calculated that our planet has a mean density about 5.5 g/cm3, more than twice the density of surface rocks like granite or sandstone. This fact meant that material inside the Earth must be much denser than surface rocks, so, Verne’s image of a Swiss cheese-like Earth could not be correct.

Once researchers realized that the interior of the Earth is denser than its surface rocks, they worked to determine how mass is distributed inside the Earth. First, they assumed that the increase in density occurred gradually, due entirely to an increase in lithostatic pressure (the pressure caused by the weight of overlying rock) with depth, for such pressure would squeeze rock together. But calculations showed that if density increased only gradually, so much mass would lie in the outer portions of the Earth that centrifugal force resulting from Earth’s spin would cause our planet to flatten into a disc. Obviously, such extreme flattening hasn’t happened, so the Earth’s mass must be concentrated toward the planet’s center. This realization led to the image of a layered Earth, with a very dense central region called the “core,” surrounded by a thick “mantle” of intermediate density. The mantle, in turn, is surrounded by a very thin skin, the relatively low density “crust.” Eventually, studies showed that the density of the core reaches 13 g/cm3, while crustal rocks have an average density of 3–6 g/cm3.

Later In the 20th C, geoscientists started to include a growing array of tools to provide further insight to the mystery of what’s inside Earth. Today’s image of the planet’s interior comes from a great variety of data sources, especially in the fields of geophysics and geochemistry. We’ll focus on seismic data as they provide key constraints on Earth’s layers.

Seismic Layering

Work by seismologists, geoscientists who study earthquakes, in the first few decades of the 20th C greatly refined our image of Earth’s interior. Seismic waves, the vibrations generated during an earthquake, travel through the Earth at velocities ranging from about 4 km/s to 13 km/s. The speed of the waves, their seismic velocity, depends on properties (including, density, compressibility, response to shear) of the material through which the waves are traveling. When waves pass from one material to another, their velocity changes abruptly, and the path of the waves bends. You can see this phenomenon by shining a flashlight beam into a pool of water; the beam bends when it crosses the interface between the two materials.

FIGURE 11.2. (a) Earthquake P-waves (compressional waves) generated at a focus travel into the Earth. We can represent the path that they follow by the use of “rays.” A “ray” is an arrow drawn perpendicular to the wave front at a point. Waves bend or “refract” as they pass downward, which happens because material properties of the Earth change with depth, so that wave velocity increases. If the change in material properties is gradual, the bending of a ray is gradual. At abrupt changes, the ray bends abruptly. P-waves bend abruptly at the mantle/core boundary. As a result, there is a P-wave shadow zone, where P-waves from a given earthquake do not reach the Earth’s surface. The presence of this shadow zone proves the existence of the core. (b) S-waves (shear waves) also travel into the Earth. But shear waves cannot pass through a liquid, and thus cannot pass through the outer core. This creates a large S-wave shadow zone. It is the presence of this shadow zone that proves the existence of liquid in the core. [14.3]

By studying the records of earthquakes worldwide, seismologists have been able to determine how long it takes seismic waves to pass through the Earth, depending on the path that the wave travels. From this data, they can calculate how seismic-wave velocity changes with depth (Figure 11.2). The data reveal that there are specific depths inside the Earth at which velocity abruptly changes and waves bend; these are called seismic discontinuities. Seismic discontinuities divide the Earth’s interior into distinct shells. Within a shell, seismic wave velocity increases gradually, and waves bend gradually, but at the boundary between shells, wave velocity changes suddenly and the waves bend. The change in seismic velocity at a seismic discontinuity can be a consequence of a compositional change, meaning a change in the identity or proportion of atoms, and/or a phase change, meaning a rearrangement of atoms to form a new mineral structure. Phase changes occur because of changes in temperature and pressure.

FIGURE 11.3. (a) The variation of compressional P-wave and shear S-wave velocities (km/s) with depth (km) in Earth (heavy black lines). The right side of the graph shows correspondence of Earth’s layering with seismic transitions. [14.4]

Figure 11.3 illustrates an average seismic velocity versus depth profile for Earth. Seismic discontinuities define layer boundaries, which led geologists to subdivide the earlier-mentioned major layers: the crust, the mantle, and the core. Today’s higher resolution observations recognize the following seismically-defined layers: the crust, the upper mantle, the transition zone (so named because it contains several small discontinuities), the lower mantle, and the outer and inner core. Let’s look at the individual layers more closely.

THE CRUST

The crust is the outermost layer of our planet. Nearly all geologic structures we find at the surface occur in rocks and sediments of the crust. Geologists distinguish between two fundamentally different types of crust on Earth: continental crust, which covers about 40% of Earth’s surface, and oceanic crust, which covers the remaining 60%.

FIGURE 11.4. Hypsometric curve of Earth, showing elevation as a function of cumulative surface area. Nearly 30% of Earth’s surface lies above modern sea-level, with the deepest oceans and highest mountains only a small fraction of the total area. Note that most continental shelf regions coincide with passive margins, underlain by stretched continental crust. [14.5]

These types of crust differ from each other in terms of both thickness and composition. The proportions of the two types of crust can be easily seen on a graph, called a hypsometric curve, which shows the percentage of Earth’s surface area as a function of elevation or depth, relative to sea level (Figure 11.4). Most continental crust lies between sea level and 1 km above sea level; mountains reach a maximum height of 8.5 km, and continental shelves (the submerged margins of continents) lie at depths of up to 600 m below sea level. Most oceanic crust lies at depths of between 3 km and 5 km below sea level. The greatest depth of the sea occurs in the Marianas Trench, whose floor lies over 11 km below sea level. The bottom of the crust is marked by a seismic discontinuity called the Moho.

Oceanic Crust

Earth’s oceanic crust is 6–10 km thick, and consists of mafic igneous rock overlain by a sedimentary blanket of varying thickness. Field studies of ophiolites (slices of oceanic crust emplaced on land by thrusting), laboratory studies of drill cores, and seismic studies indicate that oceanic crust has distinct layers. These layers, when first recognized in seismic-refraction profiles, were given the exciting names Layer 1, Layer 2a, Layer 2b, and Layer 3 (Figure 11.5).

FIGURE 11.5. (a) Columnar section with lithologic and seismologic properties of oceanic crust and upper mantle (ophiolite). (b) Samples of exhumed ocean basalt and gabbro containing peridotite xenoliths (mostly green olivine). [14.6]

Layer 3, at the base, consists of cumulate, a rock formed from mafic (magnesium- and iron-rich) minerals that were the first to crystallize in a cooling magma and then settled to the bottom of the magma chamber. The cumulate is overlain in succession by a layer of gabbro (massive, coarse-grained mafic igneous rock), a layer of basaltic sheeted dikes (dikes that intrude dikes), a layer of pillow basalt (pillow-shaped blobs extruded into sea water), and a layer of pelagic sediment (the shells of plankton and particles of clay that settled like snow out of sea water).

FIGURE 11.6. Schematic section across Earth’s active crust, labeling various features of the seafloor and the continents; scale in km. In subsequent chapters we will revisit several of the key elements. [14.5]

Self-test version.

Earth’s ocean floor can be divided into four distinct bathymetric provinces, regions that lie within a given depth range and have a characteristic type of submarine landscape, defined in the following list (Figure 11.6).

• Abyssal plains are the broad, very flat, submarine plains of the ocean that lie at depths of between 3 km and 5 km. They are covered with a layer of pelagic (deep-sea) sediment.

• Ocean ridges are long, submarine mountain ranges that rise about 2 km above the abyssal plains. Their crests, therefore, generally lie at depths of about 2–3 km. Ocean ridges are roughly symmetric relative to a central axis, along which active submarine volcanism occurs. They mark the presence of a divergent plate boundary, at which seafloor spreading occurs.

• Ocean trenches are linear submarine troughs in which water depths range from 6 to 11 km. Trenches border an active volcanic arc and define the trace of a convergent plate boundary at which subduction occurs. The volcanic arc lies on the overriding plate.

• Seamounts are submarine mountains that are not part of mid-ocean ridges. They typically occur in chains continuous along their length with a chain of oceanic islands. The island at the end of the chain may be an active volcano. A seamount originates as a hotspot volcanic island, formed above a mantle plume. When the volcano drifts off of the plume, it becomes extinct and sinks below sea level. Guyots are flat-topped seamounts. The flat top may have been formed by the erosion of a seamount as it became submerged, or it may be the relict of a coral reef that formed as the seamount became submerged.

• Submarine plateaus are broad regions where the ocean is anomalously shallow. Submarine plateaus probably form above large hot-spots.

Continental Crust

The continental crust differs in many ways from the oceanic crust. To start with, the thickness of continental crust is, on average, 4-5 times that of oceanic crust. Continental crust has an average thickness of about 40 km, while oceanic crust is on average 6km (Figure 11.7).

moho

Figure 11.7. Crustal thickness, with typical continental crust in the range 30-50km and typical oceanic crusts in the range 5-10km. Continental crust is much thicker under mountain belts and oceanic crust is much thinner at ocean ridges. Contours in km.

[USGS]

Continental crustal thickness is much more variable than oceanic crustal thickness. For example, beneath mountain belts formed where two continents collide and squash together, the crust can attain a thickness of up to 70 km, while beneath active rifts, where plates are being stretched and pulled apart, the crust thins to less than 25 km (Figure 11.8).

FIGURE 11.8. Vertical sections through continental crust of different tectonic settings, based primarily on P-wave velocities; M is Moho. Note varying thickness of highest velocity layer in the lower crust, which ranges from 20 km in arc regions to only a few kilometers in Phanerozoic orogenic belts. [14.7]

If you were to calculate the average relative abundances of elements comprising continental crust, you would find that it has, overall, a silicic to intermediate composition (that is, a composition comparable to that of granodiorite). Thus, it is less mafic, overall, than oceanic crust. Lastly, continental crust is very heterogeneous, both vertically and laterally. In contrast, oceanic crust, which all forms in the same way at ocean ridges, has the same overall structure on the planet. Specifically, continental crust consists of a great variety of different igneous, metamorphic, and sedimentary rocks formed at different times and in different tectonic settings during Earth history. The varieties of rock types have been deposited in succession, interleaved by faulting, corrugated by folding, or juxtaposed by intrusion.

In the older part of continents we find gross layering. Here, the upper continental crust has an average chemical composition that resembles that of granite (a felsic igneous rock) or granodiorite (an intermediate composition igneous rock), while the lower continental crust has an average chemical composition that resembles that of basalt (a mafic igneous rock). The boundary between upper crust and lower crust occurs at a depth of about 25 km. Contrasts in composition between upper and lower continental crust probably resulted, in part, from the differentiation of the continental crust during the Precambrian. Perhaps, partial melting of the lower crust created intermediate and silicic magmas which then rose to shallow levels before solidifying, leaving mafic rock behind. Mafic rock in the lower crust may also have formed from magma that formed by partial melting in the mantle, that rose and pooled at the base of the crust or intruded to form sills near the base of the crust. The process by which mafic rock gets added to the base of the crust is called magmatic underplating. Note that the compositional similarity between the lower continental crust and the oceanic crust does not mean that there is oceanic crust at the base of continents. The heterogeneity of the continental crust reflects its long and complicated history. The types of crust and characteristic geologic provinces are shown in Figure 11.9.

map of geologic province

Figure 11.9. Types of crust and geologic provinces. [USGS]

Most of the material forming the present continental crust was extracted from the Earth’s mantle by partial melting and subsequent rise of magma in Archean and Paleoproterozoic time, between about 3.9 Ga and 2.0 Ga. Early continental crust grew by the amalgamation of volcanic arcs formed along convergent plate boundaries, as well as of oceanic plateaus formed above hot spots. Once formed, continental crust is sufficiently buoyant that it cannot be subducted, but it can be recycled through the stages of the rock cycle. Because of its heterogeneity, it is useful to distinguish among several categories of continental crust (Figure 11.9). Differences among these categories are based on: the age and type of rock making up the crust; the time when the crust was last involved in pervasive metamorphism and deformation; and the style of tectonism that has affected the crust.

Shields. Broad regions of continents in which Precambrian rocks are presently exposed at the surface. Most of these rocks are plutonic or metamorphic, but locally, shields contain exposures of Precambrian volcanic and sedimentary rocks. Some shields (e.g., the Canadian and Siberian shields) have relatively low elevations and subdued topography, but others (e.g., the Brazilian shield) attain elevations of up to 2 km above sea level and have been deeply dissected by river erosion. Fault-bounded slices of Precambrian rocks incorporated in Phanerozoic mountain belts are not considered to be shields.
Continental platforms. Regions where Precambrian rocks are currently covered by a relatively thin veneer of unmetamorphosed and generally flat-lying latest Precambrian (Neoproterozoic) or Phanerozoic sedimentary strata. Geologists commonly refer to the rock below the sedimentary veneer as “basement” and the sedimentary veneer itself as “cover.” Cover strata of platforms is, on average, less than 2 km thick, but in intracratonic basins, it may reach a maximum thickness of 5 to 7 km. The Great Plains region of the American Midwest is part of North America’s continental platform.
Cratons. Old, relatively stable portions of continental crust. In this context, we use the word “stable” to mean that cratonic crust has not been pervasively deformed or metamorphosed during at least the past one billion years. As such, cratons include both continental platforms and pre-Neoproterozoic Precambrian shields. Because of their age, cratons tend to be relatively cool (i.e., they are regions with low heat flow), and thus quite strong. Beneath cratons, the crust has a thickness of ~40 km.
Orogenic belts. Regions of continental crust that have been involved in orogeny during the Neoproterozoic or Phanerozoic. This means that they contain igneous rocks, metamorphic rocks, and geologic structures that are younger than about one billion years old. The crust in younger orogenic belts can be either “juvenile” (meaning that it formed by extraction of new melt from the mantle at the time of orogeny), or “recycled” (meaning that it is composed of older rock that was remelted, was intensely metamorphosed, or was transformed into sediment and then lithified at the time of orogeny). Geologists subdivide younger orogenic belts into two subcategories: active or modern orogens (e.g., the European Alps, the Andes, and the Himalayas) are those in which topography and crustal thickness still reflect the consequences of mountain building processes, while inactive or ancient orogens (e.g., the Appalachians, the Caledonides, and the Brasiliano/Pan African orogens) are those in which tectonic activity ceased so long ago that erosion has exhumed deep levels of the orogen and the crust is not necessarily anomalously thick. Notably, in inactive orogens, current topography is not necessarily a relict of the uplift accompanying the original orogeny, but may be due to a much younger, independent uplift event. Because crust slowly cools through time, younger orogenic belts are warmer than cratons. Not surprisingly, heat flow in active orogens is greater than that of inactive orogens. Younger orogens generally have thicker crust than do cratons. In fact, beneath some active orogens (e.g., the Himalayan orogen), the crust reaches a thickness of 65 to 70 km.
Rifts. Regions where the continental lithosphere has been stretched and thinned. As we will see, rifting leads to the formation of normal faults, formation of deep sedimentary basins, and volcanic activity. Geologists subdivide rifts into active rifts, where stretching continues today and active faulting is occurring (e.g., the Basin and Range of the western United States, and the East African Rift), and inactive rifts, where stretching ceased long ago (e.g., Mesozoic rifts of the eastern United States). Crust in active rifts may be only 25 km thick, and is very warm—heat flow in active rifts may be many times that of cratons. Crustal thickness in inactive rifts is variable, as it reflects the history of tectonic events that the rift has endured subsequent to its original formation. Passive margins are regions of continental crust that were stretched during a rifting event that succeeded in breaking apart a supercontinent to form two pieces that are now separated by an ocean basin. When rifting ceased and sea-floor spreading began, the relict of the rift became the tectonically inactive transition between the continent and the ocean basin (hence the name “passive margin”). As time passed, the stretched continental crust of passive margins slowly subsided (sank), and became buried by a very thick wedge of sediment derived from erosion of the adjacent continent or from the growth of reefs. The top surface of this sedimentary wedge forms a broad continental shelf (such as the relatively shallow ocean bordering the east coast of North America). Crustal thickness (including the wedge of sediment) and heat flow of a passive margin depend on the age of the margin, with older margins thicker and cooler.

Age of the Crust

From these descriptions you will have noticed that age plays a key role in the various features and provinces of the crust. First, preserved continental crust is as old at ~4Ga, found in regions of Australia and Canada, for example. Second, the oldest oceanic crust is much younger, with regions of the Atlantic and western Pacific preserving rocks on the order of 200 Ma (Figure 11.10). The concept of plate tectonics provides a beautiful explanation of this contrast, where (denser) oceanic crust is consumed, but pieces of (lighter) continental crust are preserved. Comparing Figure 11.10 and 11.9 offers powerful insights into the nature of the crust, its evolution and tectonic processes in general.

map of age of last thermo-tectonic event

Figure 11.10. Age of crustal rocks of continents (bright colors) and oceans (blue hues). [USGS]

The Moho

The Moho, short for Mohorovićić discontinuity [1], is a seismic discontinuity marking the base of the crust.

Above the Moho, seismic P-waves travel at about 6 km/s, while below the Moho, their velocities are more than 8 km/s (Figure 11.8). Beneath the ocean floor, this discontinuity is very distinct and probably represents a very abrupt contact between two rock types (gabbro above, peridotite below). Beneath continents, however, the Moho is less distinct; it is a zone that is 2–6 km thick in which P-wave velocity discontinuously changes. Locally, continental Moho appears to coincide with a distinct reflector in seismic-reflection profiles, but this is not always the case. Possibly the Moho beneath continents represents different features in different locations. In some locations, it is the contact between mantle and crustal rock types, whereas elsewhere it can be a zone of sill-like mafic intrusions or underplated layers in the lower crust.

Before leaving our examination of continental and oceanic crust, we summarize their characteristics below.

Composition	Continental crust has a mean composition that is less mafic than that of oceanic crust.
Mode of formation	Continental crust is an amalgamation of rock that originally formed at volcanic arcs or hot spots, and then subsequently passes through the rock cycle. Mountain building, erosion and sedimentation, and continued volcanism add to or change continental crust. Oceanic crust all forms at mid-ocean ridges by the process of seafloor spreading.
Thickness	Continental crust ranges between 25 km and 70 km in thickness. Most oceanic crust is between 6 km and 10 km thick. Thus, continental crust is thicker than oceanic crust.
Heterogeneity	Oceanic crust can all be subdivided into the same distinct layers, worldwide. Continental crust is very heterogeneous, reflecting its complex history and the fact that different regions of continental crust formed in different ways.
Age	Continental crust is buoyant relative to the upper mantle, and thus cannot be subducted. Thus, portions of the continental crust are very old (the oldest known crust is about 4 Ga). Most oceanic crust, gets carried back into the mantle during subduction, so there is no oceanic crust on Earth older than about 200 Ma, with the exception of the oceanic crust in ophiolites that have been emplaced and preserved on continents.
Moho	The Moho at the base of the oceanic crust is very sharp, suggesting that the boundary between crust and mantle is sharp. The continental Moho tends to be less distinct.

THE MANTLE

The mantle comprises the portion of the Earth’s interior between the crust and the core. Since the top of the mantle lies at a depth of about 7–70 km, depending on location, while its base lies at a depth of 2,900 km beneath the surface, the mantle contains most of the Earth’s mass. In general, the mantle consists of very hot, but solid rock. However, at depths of about 100–200 km beneath the oceanic crust, the mantle has undergone a slight amount of partial melting. Here, about 2–4% of the rock has transformed into magma, and this magma occurs in very thin films along grain boundaries or in small pores between grains. Chemically, mantle rock has the composition of peridotite, meaning that it is an ultramafic igneous rock (i.e., it contains a very high proportion of magnesium and iron oxide, relative to silica).

Subdivision of the Mantle

The chemical composition of the mantle is roughly uniform throughout, roughly reflecting the constituents of the mineral olivine and is polymorphs (Mg-silicate). Nevertheless, the mantle can be subdivided into three distinct layers, delineated by seismic-velocity discontinuities. These discontinuities probably mark depths at which minerals making up mantle peridotite undergo abrupt phase changes. At shallow depths, mantle peridotite contains olivine, but at greater depths, where pressures are greater, the olivine lattice becomes unstable, and atoms rearrange to form a different, more compact lattices. The phase change takes place without affecting the overall chemical composition.

The three layers of the mantle, from top to bottom, are:

• Upper mantle. The shell between the Moho and a depth of about 400 km. Beneath ocean basins, the interval between about 100 km and 200 km has anomalously low seismic velocities; this interval is known as the low-velocity zone. As noted earlier, the slowness of seismic waves in the low-velocity zone may be due to the presence of partial melt; 2–4% of the rock occurs as magma in thin films along grain surfaces, or in small pores between grains. Note that the low-velocity zone constitutes only part of the upper mantle and that it probably does not exist beneath continents.

• Transition zone. The interval between a depth of 400 km and 670 km. Within this interval, we observe several abrupt jumps in seismic velocity, probably due to a succession of phase changes in mantle minerals.

• Lower mantle. The interval between a depth of 670 km and a depth of 2,900 km. Here, temperature, pressure, and seismic velocity gradually increase with depth.

As we noted above, the mantle consists entirely of solid rock, except in the low-velocity zone. But mantle rock is so hot that it deforms readily and flows very slowly (at rates of a few centimeters per year). Given the slow rate of movement, it would take a mass of rock about 100 million years to rise from the base to the top of the mantle. Because of its ability to flow, the mantle slowly undergoes convection. Convection is a mode of heat transfer during which hot material rises, while cold material sinks, analogous heating soup on a stove. Convective movement takes place in the mantle because heat from the core warms the base of the mantle. Warm rock is less dense than cool rock, and thus feels a buoyant force pushing it upwards. In other words, warmer rock has positive buoyancy when imbedded in cooler rock. In the mantle, this buoyant force exceeds the strength of plastic peridotite, and thus buoyant rock originating deep in the mantle can rise. As it does so, it pushes aside other rock in its path, like a block of wood rising through water pushes aside the water in its path. Meanwhile, cold rock, at the top of the mantle, is denser than its surroundings, has negative buoyancy and sinks, like an anchor sinking through water. Geologists still debate whether the upper mantle and lower mantle convect independently, and remain as separate chemical reservoirs, or if they mix during convection.

FIGURE 11.11. Seismic tomography of the mantle. The colors indicate differences in the velocity of seismic waves, with the region of highest velocity (coldest material) indicated by blue colors and the region of lowest velocity (hotter material) by red colors. These patterns illustrate mantle convection, in which relatively hot material rises from the core-mantle boundary region to the upper mantle, and cold material vice versa. (Courtesy J. Ritsema)

[Animation]

Because of convection, the onion-like image of the mantle that we provided above is an oversimplification. Seismic tomography (a technique for generating a three dimensional image of the Earth’s interior; Figure 11.11) suggests that the mantle is heterogeneous. Specifically, tomographic studies reveal that, at a given depth, the mantle contains regions where seismic velocities are a little faster, and regions where seismic velocities are a little slower, than an average value. These variations primarily reflect variations in temperature (waves slow down in hotter, softer mantle). An overall tomographic image shows blobs and swirls of faster-transmitting mantle interspersed with zones of slower-transmitting mantle. The slower-transmitting mantle is warmer, and less dense, and thus is rising, while the faster-transmitting mantle is colder, and denser, and thus is sinking.

Mantle Plumes

A map of active volcanoes on the Earth reveals that most volcanoes occur in chains (which, as we see later in the chapter, occur along convergent boundaries between plates), but also that some volcanoes occur in isolation. For example, several active, or recently active, volcanoes comprise the Cascades chain northwestern US, while the big island of Hawaii erupts by itself in the middle of the Pacific Ocean, far distant from any other volcano. Geologists refer to isolated volcanic sites like Hawaii as hot spots. There are about 100 hot spots currently active on the surface of the Earth.

We believe that hot-spot volcanoes form above mantle plumes, which are columns of hot mantle rising from the core-mantle boundary. According to this model, mantle plumes form because heat conducting out of the core warms the base of the mantle, creating a particularly hot, positively buoyant layer, commonly referred to as D’’. At various locations, the surface of this layer bulges up, and a dome of hot rock begins to rise diapirically. This dome evolves into a column or “plume.” Beneath a hot spot volcano, a plume has risen all the way to the base of the lithosphere. Because of the decompression that accompanies this rise, the peridotite at the top of the plume partially melts, producing basaltic magma. This magma eventually rises through the lithosphere to erupt at the Earth’s surface.

THE CORE

The core of the Earth has a diameter of about 3,481 km, which is about the size of the Moon. Because of its great density, geologists have concluded that the core consists of iron alloy. As indicated in Figure 11.2, a major seismic discontinuity divides the core into two parts. Seismic P-waves (compressional waves) can travel through the outer core and the inner core, but seismic S-waves (shear waves) travel through the inner core, but not through the outer core. Since shear waves cannot travel through a liquid, this observation means that the outer core consists of liquid (molten) iron alloy, while the inner core consists of solid iron alloy. Convective flow of iron alloy in the outer core probably generates Earth’s magnetic field, creating a potential field property that we employ for past plate reconstructions using paleomagnetism.

THE TENETS OF PLATE TECTONICS

While lying in a hospital room, recovering from wounds he received in World War I, Alfred Wegener [2] pondered the history of the Earth. He wondered why the eastern coastlines of North and South America looked like they could fit snugly against the western coastlines of Europe and Africa. He wondered why glacial tills of Late Paleozoic age crop out in India and Australia, land masses that today are close to the equator. And he wondered how a species of land-dwelling lizard could have evolved at the same time on different continents that are today separated from one another by a vast ocean. Eventually, Wegener realized that all these phenomena made sense if, in the past, the continents fit together like pieces of a jigsaw puzzle into one supercontinent, a vast Permo-Triassic landmass that he dubbed Pangea (also spelled Pangaea, both meaning “all land”).

According to Wegener, the supercontinent of Pangaea broke apart in the Mesozoic to form separate, smaller continents that have since moved to new locations on Earth’s surface. Wegener referred to this movement as continental drift. Before the breakup of Pangaea, he speculated, the Atlantic Ocean didn’t exist, so lizards could have walked from South America to Africa to Australia without getting their feet wet, and land that now lies in subtropical latitudes instead lay near the South Pole where it could have been glaciated.

Though Wegener’s proposal that continents drift across Earth’s surface seemed to explain many geologic phenomena, the idea did not gain favor with most geoscientists of the day because they could not provide an adequate explanation of why continents moved. Such an explanation did not appear until early 1960s, when Harry Hess [3] circulated an influential manuscript in which he proposed a process known as seafloor spreading. During seafloor spreading, new ocean floor forms along the axis of a submarine mountain range called an ocean ridge. Once formed, the new ocean floor moves away from the ridge axis. Two continents drift apart when seafloor spreading causes the ocean basin between them to grow wider.

Soon after, researchers from around the world rushed to explore the implications of the seafloor spreading hypothesis and to reinterpret the phenomena of continental drift. The result of this work led to a broad set of ideas, which together comprise the unifying Theory of Plate Tectonics (or, simply, plate tectonics). According to this theory, the lithosphere, Earth’s relatively rigid outer shell, consists of discrete pieces, called lithospheric plates (or, tectonic plates), which move relative to one another. Continental drift and seafloor spreading are manifestations of such plate motion.

Plate tectonics is a geotectonic theory. It is a comprehensive set of ideas that explains the development of regional geologic features, such as the distinction between oceans and continents, the origin of mountain belts, and the distribution of earthquakes and volcanoes, the formation and distribution of main rock types. Acceptance of plate tectonics represented a revolution in geology, for it led to the conclusion that Earth is mobile. Prior to plate tectonics, most geologists had a fixed Earth view, in which continents were fixed in position through geologic time. In this context, mountain building was viewed as a consequence of predominantly vertical motions. Pre-plate tectonics ideas to explain mountain building included: (1) the geosyncline hypothesis, which stated that mountain belts formed out of the deep, elongate sedimentary basins (known as “geosynclines”) that formed along the margins of continents. According to this hypothesis, mountain building happened when the floor of a geosyncline sank deep enough for sediment to melt; the resulting magma rose and in the process deformed and metamorphosed surrounding rock; and, (2) the contracting Earth hypothesis, which stated that mountains formed when the Earth cooled, shrank and wrinkled, much like a baked apple removed from the oven. Both ideas have been thoroughly discredited. Today, the map of our planet constantly changes, though slowly.

This narrated module (IRIS) offers a 6min informative overview of the basic tenets of plate tectonics.

FIGURE 11.12. The seven major and several minor plates of today’s Earth, and their boundaries; in Molweide projection. Major plates are: Pacific, North American, South American, Eurasian, African, Indo-Australian and Antarctic plates; smaller plates include: Juan de Fuca, Caribbean, Cocos, Nazca, Scotia, Arabian, Philippine plates. Ocean ridges and transfer faults (heavy lines) and trenches (lines with teeth on the overriding plate) mark the plate boundaries. In a few places, plate boundaries are ill-defined, such as the Indian and Australian plate boundary (here Indo-Australian Plate) and the North American and Eurasian plate boundary in Siberia. [Modified from Scotese's PALEOMAP project]

Self test version of today's plates. Identify the major and minor plates, and the nature of their boundaries.

Instead of sheets on a flat table, a plate must be viewed geometrically as a cap on the surface of a sphere, which has significant implications for global displacements (see below). The border between two adjacent plates is a plate boundary. During plate movement, the plate interior (the region away from the plate boundary) stays relatively coherent and undeformed; such plate rigidity is one of the tenets of the theory. Instead, most plate movement is accommodated by deformation along plate boundaries, and this deformation generates most earthquakes. A map of earthquake epicenters defines seismic belts, and these belts define the locations of today’s eight major plates and several microplates (Figure 11.12).

FIGURE 11.13. The basic types of plate boundaries: divergent, transform and convergent boundaries. [14.14]

Geoscientists distinguish three types of plate boundaries (Figure 14.14):

At divergent plate boundaries, defined by mid-ocean ridges (also called oceanic ridges, because not all occur in the middle of an ocean), two plates move apart as a consequence of seafloor spreading. Thus, these boundaries are also called spreading centers. The process of seafloor spreading produces new oceanic lithosphere.
At convergent plate boundaries, one oceanic plate sinks into the mantle beneath an overriding plate, which can be either a continental or an oceanic plate. During this process, which is called subduction, an existing oceanic plate gradually disappears. Thus, convergent plate boundaries are also called destructive boundaries or consuming boundaries. Volatile elements (water and carbon dioxide) released from the subducted plate trigger melting in the overlying asthenosphere. The resulting magma rises and erupts in a chain of volcanoes, called a volcanic arc, along the edge of the overriding plate. “Arcs” are so named because many, though not all, are curved in map view. The actual plate boundary at a convergent boundary is delineated by a deep ocean trench. Convergence direction is not necessarily perpendicular to the trench. Where convergence occurs at an angle of less than 90° to the trend of the trench, the movement is called oblique convergence.
At transform plate boundaries, one plate slides past another along a strike-slip fault. Since no new plate is created and no old plate is consumed along a transform, such a boundary can also be called a “conservative boundary.” Transform plate boundaries can occur either in continental or oceanic lithosphere. At some transform boundaries, there is a slight component of convergence, leading to compressive stress. Such boundaries are called transpressional boundaries. Similarly, at some transform boundaries there is a slight component of divergence, leading to tensile stress. Such boundaries are called transtensional boundaries.

Not all plate boundaries are sharp lines. Some, such as the boundary that occurs between India and Asia, and the boundary between parts of western North America and the Pacific Plate, define diffuse plate boundaries. These typically occur where plate boundaries lie within continental crust, for the quartz-rich continental crust is relatively weak and continental crust contains many preexisting faults, so the deformation is not confined to a narrow zone. Some seismic activity and deformation does occur entirely away from a plate boundary. Such activity, called plate-interior deformation, probably indicates the presence of a particularly weak fault zone within a plate, capable of moving in response to the ambient stress state within a plate. Seismic plate-interior deformation currently occurs along the New Madrid Fault System in the central US.

FIGURE 11.14. (a) Schematic cross section of a collision zone, where two buoyant continents converge, creating a broad belt of deformation and crustal thickening; and (b) a rift, where stretching of the crust causes thinning and normal faulting. [14.15]

In addition to the plate boundaries just discussed, geologists recognize two other locations where movement of lithosphere creates structures.

At collision zones, two buoyant blocks of crust converge (Figure 11.14a). The buoyant blocks, which may consist of continental crust, island arcs, or oceanic plateaus, are too buoyant to be subducted, so collisions result in broad belts of deformation, metamorphism, and crustal thickening. At the end of a collision, two continents that were once separate have become stuck together to form one continuous continent; the boundary here is called a suture. Large continents formed when several smaller continents have sutured together are called supercontinents.
At a rift, an existing continent stretches and starts to split apart (Figure 11.14b). At a successful rift, the continent splits in two and a new mid-ocean ridge forms. The stretched continental crust along the boundary of a successful rift evolves into a passive continental margin. At an unsuccessful rift, rifting stops before the split is complete so the rift remains as a permanent scar in the crust. It is usually marked by an elongate trough, bordered by normal faults and filled with thick sediment and/or volcanic rock.

As shown in Figure 11.12, the Earth currently has seven major plates (Pacific, North American, South American, Eurasian, African, Indo-Australian, and Antarctic plates) and several smaller plates (e.g., Cocos, Nazca, Arabian, and Philippine plates). A plate can consist entirely of oceanic lithosphere (such as the Pacific and Nazca plates), but most plates consist of both oceanic and continental lithosphere. For example, the North American Plate consists of the continent of North America and the western half of the Atlantic Ocean, and the Indo-Australian plate consists of these two continents and part of the Indian Ocean. Continental margins are not necessarily plate boundaries, so, for this reason, we make the important distinction between active continental margins, which are plate boundaries, and passive continental margins, which are not plate boundaries. The western margin of Africa is a passive margin, but the western margin of South America is an active margin. As noted earlier, passive margins form from stretched continental lithosphere left by the rifting that led to the successful formation of a new ocean ridge.

Continents on either side of an ocean ridge move apart as seafloor spreading causes the intervening ocean basin to grow. Continents separated by a convergent boundary move together as subduction consumes intervening seafloor. Thus, plate motions cause the map of Earth’s surface to constantly change through time, so Wegener’s hypothesis of continental drift (s.l.), represented by moving land masses, does indeed occur (Figure 11.15).

FIGURE 11.15. A sequence of labeled maps showing the position of Earth's continental elements since the Early Mesozoic breakup of Pangaea (Molweide projection) and an animation from 540Ma to today. [Courtesy Scotese's PALEOMAP project]

Insights from Earthquakes

Patterns of earthquake epicenters mark the location of today’s plate boundaries. The nature of earthquake displacements as obtained from earthquake analysis also gives us the type of plate boundary. Plotting epicenters on a map of the world highlights the predictive power of plate tectonic theory for future earthquake occurrences around the world (Figure 11.16)

FIGURE 11.16. A map of >21,000 earthquake epicenters (dots) in 1998 outlines the locations of plate boundaries. A small fraction, the exception, are intraplate earthquakes that occur in plate interiors. Earthquake magnitude is reflected in dot size. Earthquake depth is indicated by color: shallow (0-70km is yellow; intermediate (71-300km) is orange; deep (301-700km) is red. (from IRIS; https://goo.gl/bOujpp) [14.13]

Locating earthquakes in the third dimension (depth) produces another, inclined belt, that we call a Wadati-Benioff zone [4], which reaches a maximum depth at around 670 km. This distribution of earthquakes defines the location of the subducting plate with depth (Figure 11.17).

FIGURE 11.17. (a) Trenches (heavy lines), volcanoes (black dots) and backarc basins related to subduction in the western Pacific. The depth to the subducted slab (Wadati-Benioff zone) is shown by contour lines, given in multiples of 50 km (e.g., 2 means 100 km depth below surface.) (b) Cross section showing earthquake foci defining the moderately dipping Wadati-Benioff zone of the northern Izu-Bonin arc, and earthquake foci defining a steeply dipping Wadati-Benioff zone of the northern Mariana Arc. T = location of trench; V = location of volcanic arc. [17.6]

Based on the shape of the Wadati-Benioff zone, researchers find that not all subducted plates dip at the same angle. In fact, dips vary from nearly 0°, meaning that the slab shears along the base of the overriding slab, to 90°, meaning that it plunges straight down into the mantle. Subducted-slab dip may be controlled, in part, by the age of the subducting lithosphere, for older oceanic plate is denser and may sink more rapidly. It may also be controlled by convergence rate, the horizontal rate at which plates are converging across the trench, for if we assume that sinking velocity is constant, an increase in convergence velocity decreases the dip of the subducting plate. The angle may also be affected by the flow direction and velocity of the asthenosphere into which the lithospheric plate sinks.

Because the lithosphere is cooler than asthenosphere, the subducting plate perturbs the thermal structure of the mantle. The internal part of the downgoing plate remains relatively cool down to significant depths because rock has such low thermal conductivity. Under the pressure and temperature conditions found in the subducting plate, basalt of the oceanic crust undergoes a phase transition to become a much denser rock called eclogite; formation of eclogite increase the slab-pull force.

The type of stress associated with earthquakes changes character with depth along the Wadati-Benioff zone. Beneath the outer swell, earthquakes result from tension caused by plate bending, whereas in the region beneath the accretionary wedge, earthquakes result from compression; thrust movements are due to shear between the overriding and down-going slab. Huge, destructive earthquakes, such as the 1964 “Good Friday” earthquake of southern Alaska, result from ruptures in this zone. At depths of about 150–300 km, earthquakes of the Wadati-Benioff zone occur in a tensional stress field. Perhaps slab pull by the deepest part of the subducting plate stretches the plate in this interval. At deep levels, earthquakes of the Wadati-Benioff zone indicate development of compression, perhaps caused by shear between the deep down-going plate and the asthenosphere. Seismologists do not well understand why deep-focus earthquakes of the Wadati-Benioff zone can occur, because at great depths, the down-going slab should be warm enough to be ductile. Some researchers suggest that deep-focus earthquakes happen when sudden mineralogical phase transition or sudden dehydration reactions take place in rock comprising the down-going plate, and that these cause an abrupt change in the volume of the rock; this change generates vibrations. The deepest earthquakes of the Wadati-Benioff zone occur near the boundary between the seismically defined transition zone of the mantle and the lower mantle.

FIGURE 11.18. Schematic cross section of Earth illustrating the concept of a slab graveyard in which masses of (denser, colder) subducted oceanic lithosphere may accumulate near the base of the mantle. [17.9]

Earthquakes from depths greater than 670 km have not been detected. But the deepest earthquakes do not necessarily define the greatest depth to which a down-going slab sinks. Seismic tomography studies indicate that down-going plates, because they are relatively cool, show up as bands of anomalously fast velocity. Some bands continue downwards into the lower mantle, suggesting that the down-going slab flows downwards into the lower mantle. Subducted plates may eventually sink almost to the base of the mantle, accumulating in slab graveyards at the core-mantle boundary (Figure 11.18). If this image is correct, then the base of the Wadati-Benioff zone does not mark the base of subducted slabs, but merely the depth at which earthquakes no longer occur because fracturing and/or phase changes that produce seismic energy in slabs no longer occur.

KINEMATICS OF PLATE TECTONICS

When we talk of plate kinematics, we are referring to a description of the rates and directions of plate motion on the surface of the Earth. Description of plate motion is essentially a geometric exercise, made somewhat complex because the motion takes place on the surface of a sphere. Thus, our description must use the tools of spherical geometry, and to do this, we make three assumptions:

We assume that the Earth is a sphere. In reality, the Earth is slightly flattened at the poles (the Earth’s radius at the poles is about 20 km less than it is at the equator), but this flattening is not sufficiently large for us to worry about.
We assume that the circumference of the Earth remains constant through time. This assumption implies that the Earth neither expands nor contracts, and thus that the amount of seafloor spreading worldwide is equivalent to the amount of subduction worldwide. Note that this assumption does not require that the rate of subduction on one side of a specific plate be the same as the rate of spreading on the other side of that plate, but only that growth and consumption of plates are balanced for all plates combined over the whole surface of the Earth.
We assume that plates are internally rigid, meaning that all motion takes place at plate boundaries. This assumption, as noted earlier, is not completely valid because continental plates do deform internally, but this error for plate motion calculations is less than a few percent.

FIGURE 11.19. Plate kinematics. (a) A sequence of map-views illustrating the movement of two continents with seafloor spreading between them. Both continents are drifting east, in an absolute reference frame (black line on the left). At the same time, B is moving east relative to A, which is equivalent to saying that A is moving west relative to B (red line). (b) Comparing velocities of plates A and B in different reference frames. First we consider relative velocities (red arrows). Plate A is moving west at 2 cm/y relative to the mid-ocean ridge (MOR), and plate B is moving east at 2 cm/y relative to the MOR. Thus, plate B moves east at 4 cm/y relative to fixed plate A, and plate A moves at 4 cm/y west relative to fixed plate B . However, considering absolute velocities (i.e., velocities relative to a fixed point in the mantle), plate A moves east at 2 cm/y relative to a fixed point, while plate B moves east at 6 cm/y relative to a fixed point (black arrows). Note that the relative motion of plate B with respect to plate A occurs because plate B’s absolute velocity to east is faster than that of plate A. [14.17]

We use two different reference frames to describe plate motion (Figure 11.19), which you are encouraged to follow with, say, two coasters on a table. When using the absolute reference frame, we describe plate motions with respect to a fixed point in the Earth’s interior. When using the relative reference frame, we describe the motion of one plate with respect to another. To understand this distinction, imagine that you are describing the motion of two cars cruising down the highway. If we say that Car A travels at 60 km/h, and Car B travels at 40 km/h, we are specifying the absolute velocity of the cars relative to a fixed point on the ground. However, if we say that Car A travels 20 km/h faster than Car B, we are specifying the relative velocity of Car A with respect to Car B. We first look at how we specify absolute plate velocity and then we will examine relative plate velocity.

Absolute Plate Velocities

We mentioned earlier that not all volcanoes occur along plate boundaries. Some, called hot-spot volcanoes, erupt independently of plate-boundary activity; they form above a deeper mantle plume. Since deep mantle plumes are independent of plate boundaries, they can be used as the fixed reference points for calculating absolute plate velocities. The hot-spot frame of reference gives a reasonable approximation of absolute plate velocity, but, in reality, it’s not perfect because hot spots actually do move with respect to one another. Nevertheless, the velocity of hot-spot movement is an order of magnitude less than the velocity of plate movement, so the error in absolute plate motions based on the hot-spot reference frame is only a few percent.

Figure 11.20 compares the relative and absolute velocities of plates. You will notice that absolute plate rates today range from less than 1 cm/y (for the Antarctic and African Plates) to about 10 cm/y (for the Pacific Plate). North America and South America are moving west at about 2–3 cm/y. For comparison, these rates are about the same at which your fingernails grow. This may seem slow, but remember that a velocity of 2 cm/y yields a displacement of 2000 km in 100 million years! This means that there is plenty of time in geologic history for large oceans to open and close several times.

FIGURE 11.20. Directions and rates in mm/y of plate motion. Relative motions of the major plates are given for selected points along their boundaries by pairs of inward-pointing (convergent) or outward-pointing (divergent) arrows. Numbers next to these arrows give relative velocity in cm/y. [14.19]

Relative Plate Velocities

The motion of plates with respect to one another is defined by fixing the position of one of the plates. For example, if we describe the motion of plate A with respect to plate B, we fix plate B on the surface of the Earth and see how plate A moves (Figure 11.19). Recall that Earth is a sphere, so, in plate kinematic calculations, this movement is described by rotation with a specified angular velocity around an imaginary rotation axis that passes through the center of Earth. The intersection between this rotation axis and the surface of the Earth is called an Euler pole (Figure 11.20a–c).

FIGURE 11.20. The motion of plates on a spherical Earth. (a) Latitudes and longitudes around the Euler (rotation) pole (E) are small circles and great circles, respectively. The plate motion from time 1 to time 2 is a rotation (ω) around the Euler pole. Linear velocity (v) is measured on the surface of Earth. (b) Transform faults describe small circles (latitudes of rotation) relative to the Euler pole. (c) The relation of transforms relative to the Euler pole in a projection looking down the pole. Notice that (linear) spreading velocity and rate of slip along transform faults change with distance from the Euler pole. [14.20]

Euler poles are geometric elements that are not related to the Earth’s geographic poles (the points where the Earth’s spin axis intersects the surface), nor are they related to the Earth’s magnetic poles (the points where the Earth’s internal dipole intersects the surface). We distinguish between two types of Euler poles. An instantaneous Euler pole is used to describe relative motion between two plates at an instant in geologic time, whereas a finite Euler pole describe total relative motion over a long period of geologic time. For example, the present-day instantaneous Euler pole describing the motion of North America with respect to Africa determine how fast Chicago is moving away from Casablanca today. We can calculate a single finite Euler pole to describe the motion between these two locations over the past 80 million years, even if the instantaneous Euler pole changed at several times during this interval. In Supplements we examine vector analysis and angular rotations to further explore plate displacements and other implications.

Until the end of the 20th C, relative plate motion could not be obtained by direct observation. But today, the global positioning system (GPS), which uses signals from an array of satellites to determine the location of a point on Earth’s, makes this possible. By setting up a permanent network of GPS stations, we are able to define the location of a point within a few millimeters, and with this accuracy, plate movements over a period of a few months to years can now be detected.

Triple Junctions

At this point, you should have a clear idea of what we mean by a plate boundary between plates. We represent the intersection of a plate boundary with the surface of the Earth by a line on a map. The point where three such plate boundaries meet is called a triple junction. We name specific types of triple junctions by listing the plate boundaries that intersect. For example, at a ridge-ridge-ridge triple junction, three divergent boundaries intersect (e.g., in the southern South Atlantic and in the southern Indian Ocean); at a ridge-trench-transform triple junction, a divergent boundary, a convergent boundary and a transform boundary intersect (e.g., off the west coast of northern California).

FIGURE 11.21. Geometry of triple junctions. (a) This ridge-trench-transform triple junction is stable. (b) With time the ridge-trench-transform triple junction location changes (T to T’), but the geometry stays the same. [14.22]

We further distinguish between stable and unstable triple junctions. The basic configuration of a stable triple junction can exist for a long time, even though the location of the triple junction changes. For example, the ridge-trench-transform junction in Figure 11.21 represent a stable triple junction, because, while the location of the triple junction migrates to the southeast with time (point T to T’), the plate geometry remains the same. In contrast, the geometry of an unstable triple junction changes to create a new arrangement of plate boundaries. The migration of a triple junction along a plate boundary can lead to transformation of a segment of a plate boundary from one type of boundary to another. Such a change occurred along the coast of California, an active continental margin, during the Cenozoic. Through most of the Mesozoic, and into the Early Cenozoic, the western North American margin was a convergent plate boundary. Toward the end of this time interval, convergence occurred between the North American and Farallon plates. At around 30 Ma, the Farallon-Pacific ridge (the divergent boundary between the Farallon and Pacific plates) began to be subducted. When this occurred, the Pacific Plate came into contact with the North American Plate and two triple junctions formed, one moving north-northwest, and the other moving south-southeast. The margin between the triple junctions changed from being a convergent boundary to a transform boundary, the San Andreas Fault.

MECHANICS OF PLATE TECTONICS

Alfred Wegener was unable to convince the geologic community that continental drift happens, because he could not explain how it happens. The question of what exactly drives the plates remains debated to this day, but the theory is unquestioned. In the years immediately following the proposal of plate tectonics, many geoscientists tacitly accepted a convection-cell model, which states that convection- driven flow in the mantle is driving the plates. In this model, plates are carried along on the back of flowing asthenosphere, which was thought to circulate in simple elliptical (in cross section) paths; upwelling (upward flow) of hot asthenosphere presumably occurred at ocean ridges, while downwelling (downward flow) of hot asthenosphere occurred at the margins of oceans or at subduction zones. In this model, the flowing asthenosphere exerts basal drag, a shear stress, on the base of the plate, which is sufficient to move the plate. This image of plate motion, however, is now discarded. While it is clear from numerical calculations and seismic tomography that the mantle does convect, it is impossible to devise a global geometry of convection cells that explains the observed geometry of plate boundaries that exist on Earth.

FIGURE 11.22. Generalized diagram illustrating the forces acting on a tectonic plate and representative densities of mantle, shallow (basaltic) plate and deep (eclogitic) plate. [14.23]

Plate-driving Forces

Modern calculations showed that two forces, ridge push and slab pull, play the major role in driving plates (Figure 11.22):

Ridge push is the outward-directed force that pushes plates away from the axis of a mid-ocean ridge. It exists because oceanic lithosphere is higher along ocean ridges than it is in the abyssal plains at a distance away from the ridge. In Earth’s gravity field, the difference in elevation means that the lithosphere along the ridge has more gravitational potential energy than the lithosphere of the abyssal plain, and this energy provides an outward push perpendicular to the ridge axis. To picture this push, imagine a thin pool of molasses on a table top. If you pour more molasses onto the center of the sheet, a mound of molasses builds up in the center. The new molasses flows outward and pushes the molasses that was already on the table in front if it, causing the diameter of the molasses sheet to increase.
Slab pull is the force that pulls lithosphere into a convergent margin. It exists because old, cold ocean lithosphere is negatively buoyant relative to the underlying asthenosphere, so if given a chance, the oceanic lithosphere sinks downwards. Once the cold subducting plate (also called a subducted slab) descends into the mantle, phase transformations change basalt to much denser eclogite, so a subducted plate is denser than a plate at the Earth’s surface. Thus, like a sinking anchor, a subducted slab pulls the rest of the plate down with it. To picture the process, imaging that you smooth a sheet of wet tissue paper onto the ceiling. If you peel one end of the paper off, and let go, the paper slowly moves down and peels the rest of the paper off with it, until the whole sheet eventually separates from the ceiling and falls.

Let’s recap by considering all the forces that act on a plate (Figure 11.22). Ridge-push forces drive plates away from mid-ocean ridges, and slab-pull forces drive them down subduction zones. Thus, plate motion is, to some extent, a passive phenomenon, in that it is the gravitational potential energy of the plate itself that causes a plate to move. Does basal drag due to mantle convection play a role at all in driving plates? Probably yes, but not in the simple way that was first envisioned. Convection in the mantle, according to seismic tomography, seems to be accommodated globally by a few upwelling zones and a few downwelling zones. These zones do not exactly correspond to the present configuration of ridges and trenches. The flow of asthenosphere due to convection probably does create some basal drag at the base of plates, but the basal drag force can assist or retard motion. Specifically, where the asthenosphere flows in the same direction as the plate motion caused by ridge push and/or slab pull, basal drag may accelerate the motion. In contrast, where the asthenosphere flows in a direction opposite to the plate motion caused by ridge push and/or slab pull, it may retard the motion. And if the asthenosphere flows at an angle to the plate motion caused by ridge push and/or slab pull, it may change the direction of motion. Calculations show that the basal-drag force caused by asthenospheric flow is less than the force of ridge push, which, in turn, is less than the slab pull force. Lastly, the forces that drive plates are resisted by frictional forces between plates, at plate boundaries. Whereas convection of the asthenosphere alone does not drive plate motion, mobility of the asthenosphere is needed to make plate motion possible. If the asthenosphere could not flow up at ocean ridges to fill space created by seafloor spreading, and if the asthenosphere could not move out of the way of subducting slabs, then buoyancy forces (negative or positive) would not be sufficient to cause plates to move. Using a car analogy for plate tectonics, the mantle is the (heat) engine of the automobile, but the plate sits in the driver's seat. Have a look at the IRIS animation on "What drives plate tectonics?" (~7min).

PLATE TECTONIC CYCLES

We close this overview chapter by looking at two organizing patterns in plate tectonics, the Wilson Cycle and the Supercontinent Cycle.

Figure 11.23. The Wilson Cycle. a-b) A continent rifts, such that the crust stretches, faults and subsides. c) Seafloor spreading begins, forming a new ocean basin. d) The ocean widens and is flanked by passive margins. e) Subduction of oceanic lithosphere begins on one of the passive margins, closing the ocean basin. (f-g) The ocean basin is destroyed by continental collision. At some later time continental rifting begins again and the process repeats.

Wilson Cycle

Soon after the basic tenets of seafloor spreading were established, Tuzo Wilson [5], noted that formation of the Appalachian-Caledonide Orogen in eastern North America and Western Europe involved closure of a vanished ocean. Based on various data, he reasoned that there must have been another ocean (not the present-day Atlantic) before the formation of Pangaea separating these continents, the Proto-Atlantic (or Iapetus) Ocean. He envisioned a cycle of tectonic activity in which an ocean basin opens by rifting, grows by seafloor spreading, closes by subduction, and disappears during continent-continent collision, and that more than one such cycle has happened during Earth history. The successive stages of rifting, seafloor spreading, convergence, collision, rifting, recorded in a single mountain range came to be called the Wilson cycle (Figure 11.23). Because of the Wilson cycle, we do not find oceanic lithosphere older than about 200 Ma on Earth. It has all been subducted. Very old continental crust, however, remains, because it’s too buoyant to be subducted, which explains why we can still find Archean rocks on today’s continents.

FIGURE 11.24. Stages in the supercontinent cycle. (a) Continents move relative to one another, but gradually aggregate over a mantle downwelling zone. (b) While the supercontinent exists, large-scale convection in the mantle reorganizes. (c) Upwelling begins beneath the supercontinent and weakens it. Eventually, rifting occurs and the supercontinent breaks up, forming smaller continents that drift apart. [14.24]

Supercontinent Cycle

Subsequently, geologists realized that Wilson cycles were part of a global succession of events leading to formation of a supercontinent, breakup of a supercontinent, dispersal of continents, and reassembly of continents into a new supercontinent. This succession of events has come to be known as the supercontinent cycle (Figure 11.24). At various times in the past, continental movements and collisions have produced supercontinents, which lasted for tens of millions of years before eventually rifting apart. The geologic record shows that the supercontinent Pangaea had formed by the end of the Paleozoic (~250 Ma), only to disperse in the Mesozoic. Similarly, the supercontinent Rodinia (Figure 11.25) had formed by the end of the Mesoproterozoic (~1.1 Ga), only to disperse at about 900 Ma to form a new supercontinent, Pannotia (~600 Ma), which itself broke up at the end of the Neoproterozoic. And there is growing evidence that supercontinents formed even earlier in Earth history, such as Nuna at the end of the Paleoproterozoic (~1.8Ga). Thus, it seems that supercontinents have formed, broken up, formed, and broken up at 100’s m.y. intervals.

FIGURE 11.25. Reconstruction of the Mesoproterozoic Rodinia supercontinent, showing lines of subsequent (Neoproterozoic) opening of the Pacific and Iapetus Ocean Basins. Rodinia is an aggregate of cratons cemented by ~1.1 Ga orogens, including the Grenville Orogen of North America and correlatives elsewhere. [22.6.3]

Given their recurring formation, it has been suggested that supercontinents may be related to long-term convection patterns in the mantle. In these models, relative motion between plates at any given time is controlled by plate forces, but over longer periods of time. Continents tend to accumulate over a zone of major mantle downwelling to form a supercontinent (Figure 11.24). However, supercontinents don’t last forever; once a supercontinent forms, the thermal structure of the mantle beneath it changes. This change happens because a supercontinent acts like a giant insulator that does not allow heat to escape from the mantle. Over 80% of the Earth’s internal heat escapes at mid-ocean ridges, where seawater circulating through the hot crust and upper mantle cools the lithosphere much like a coolant cools an automobile engine; however, within the area of a supercontinent, there are no ridges, so heat cannot easily be lost. As a consequence, the mantle beneath the supercontinent eventually heats up, and can no longer be a region of downwelling. When this happens, a new downwelling zone develops elsewhere, and hot asthenosphere must begin to upwell beneath the supercontinent. Such upwelling causes the continental lithosphere of the supercontinent to heat up and weaken, and it rifts apart into smaller continents separated by new oceans. If the supercontinent-cycle model works, then we can say that Earth is currently in the dispersal stage of this cycle. During the next 200 million years, the assembly stage will begin and the floors of the Atlantic and Pacific Oceans will be subducted. After a series of collisions, the continents may once again coalesce to form a new supercontinent that has been called, and why not, Amasia.

Back to Plate Tectonics Contents

Back to PSG&T Home

v1.3; last update: 15-Sep-2023

[1] Named for Andrija Mohorovićićc, the Croatian seismologist who discovered it in 1909.

[2] A German meteorologist and geologist (1880–1930).

[3] An American geologist (1906-1969). Seafloor spreading is also credited to Robert Dietz (1914-1995).

[4] Named after K. Wadati (Japan) and H. Benioff (USA) working independently, several decades apart.

[5] Canadian geologist (1908-1993).