Mantle Dynamics and Plate Kinematics

Carlo Doglioni

La Sapienza University, Rome, Italy

Roberto Sabadini

University of Milan, Italy

Keywords: lithosphere, asthenosphere, mantle, core, transition zone, viscosity, rheology, gravity field, Earth’s rotation, westward drift, plate kinematics, geodynamics, Pleistocene deglaciation, ice mass instability, Antarctica, Greenland, decollement, subduction zone, rollback, back-arc basin, orogens, foredeeps, foreland monocline, rift zones, transform faults


1. Introduction

2. Techniques to Sample the Interior of the Earth

3. Seismic Tomography

4. Changes in Earth’s Gravity Field and their Implications for Mantle Dynamics

5. Dynamic Structure of the Mantle and Ice Loss in Antarctica and Greenland

6. How the Rheology of the Mantle Impacts the Style of Subduction

7. Plate Kinematics

8. The Hot Spot Reference Frame and the Westward Drift of the Lithosphere

9. Plate Boundaries

10. Plate Kinematics versus Mantle Dynamics


Related Chapters



Biographical Sketches


Plates move at the surface of Earth, but we still do not know what energy source accounts for plate tectonics. Mantle tomography, studies on the viscosity of Earth’s interior, and geological, geophysical, and geodetic analysis of plate tectonics are rapidly contributing to an understanding of why plates describe a sinusoidal flow and why there is a westward delay of the lithosphere relative to the mantle, creating strong asymmetries in the structure of subduction zones and rift zones. The main energy seems to come from Earth’s cooling and associated mantle convection. Earth’s rotation however seems to contribute as well, both in terms of direction of plate motions, and possibly also in terms of energy.

1. Introduction   

Life on Earth is dependent on its atmosphere, which formed mainly as a result of mantle degassing through volcanoes and other forms. Volcanism on Earth is primarily a consequence of plate tectonics, and therefore the motion of crustal plates is a foundation of life on Earth. Plate tectonics lead to disasters such as earthquakes, volcanic eruptions, and landslides, but are also the engine for natural resources, from hydrocarbons to geothermy, mineral resources, water, and the spectacular landscapes that enhance our lives. In other words, we cannot live without plate tectonics, and it is very helpful to better understand the mechanisms and the structures associated with the movements of the lithosphere.

The lithosphere is subdivided in a number of plates; a plate is an element of the lithosphere characterized by its own independent velocity relative to the neighboring sections of the lithosphere. These differential velocities among plates generate plate tectonics. Approaching or separating plates are controlled by their relationship with the underlying mantle. There is a close link between what happens in Earth’s mantle and what is going on at the surface. In this article we briefly describe what is understood about mantle dynamics and plate kinematics, and link the two.

2. Techniques to Sample the Interior of the Earth   

Nowadays, together with the study of magmatic rocks (that is, rocks formed deep in the Earth, but now found in its crust), satellite geodesy and seismic tomography are major tools for sampling the interior of Earth, to retrieve its innermost structure in terms of its density, and its elastic and rheological properties. "Rheology" covers a wide field of studies related to the creep and flow properties of metals, ice, or in this case, Earth’s mantle. The term "rheology" can be used to indicate not only the field of study but also the property of flowing and creeping of materials, so that it is possible to speak of the rheology of ice or the rheology of the mantle. Rocks have the ability to "flow" under the impact of the stresses that originate from geodynamic processes. The major region of the Earth where geodynamic processes originate is the mantle, the portion of our planet between the rigid outer lithosphere, which is about 100 km thick, and the core, at a depth of 2890 km. Figure 1 is a sketch showing schematically the interior of our planet, with the mantle characterized by the fundamental property of creep, which means that the mantle rocks are able to flow when subject to shear stresses applied for time intervals typical of geological processes, from thousands to million years. In Figure 1, the two layers embedded between the elastic lithosphere and the core denotes the mantle. The mantle itself is subdivided into the upper and lower mantle, with the transition zone located at a depth of 670 km.

Figure 1. Model of Earth. The lithosphere is both oceanic and continental, and includes the crust. It is the outermost layer, and it behaves mainly as an elastic body. The upper and lower mantle are viscoelastic, which means that mantle rocks creep under the action of long-standing shear stresses, while the external core is an inviscid fluid. Variations in density in the mantle are inferred by seismic tomography, and they refer to variations relative to similar depths. Boundaries between the different layers represent changes in density and viscosity. Note that the different shells are interpreted to be shear zones. In particular the lithospheric movement is delayed westward relative to the mantle, determining a relative eastward mantle flow.

A key issue in geodynamics is the determination of the viscosity of the upper and lower mantle; the viscosity is higher in the lower mantle. The importance of the viscosity stands on the fact that the dynamics of the mantle and the kinematics of the plates are ultimately controlled by this parameter, which characterizes the flow properties of Earth’s interior. The innermost portion of Earth is the core, as indicated in Figure 1, which differs substantially from the mantle in terms of rheological properties: the external core is in fact an inviscid fluid, whereas the internal core is solid and is the densest part of the planet. The outermost layer of Earth is called the "lithosphere," and in contrast to the mantle, its viscosity is so high that it can be assumed to behave as an elastic body. The five-layer model in Figure 1 (lithosphere, upper mantle, lower mantle, external core, and internal core) shows all the fundamental ingredients of Earth in terms of its rheological and dynamic properties.

3. Seismic Tomography   

Seismic tomography is the technique of inverting seismological data to retrieve a three-dimensional image of the anomalies in seismic wave velocity within the media they cross. It is based on the physical phenomenon that the velocities of seismic waves (that is, waves triggered by earthquakes) are subject to change as they cross regions of Earth’s mantle with different densities and elastic properties. The availability of a large amount of data from the well-recorded global seismicity thus makes it possible today to retrieve a picture of the structure of Earth’s mantle, in terms of thermal and compositional anomalies.

Both surface wave and body waves can be used to detail the image of Earth’s interior. Body waves are those that have the ability to cross the whole mantle, while surface ones propagate in proximity to the interfaces where the density or elastic properties of the Earth are discontinuous. The surface waves thus provide information on Earth’s structure in the upper mantle, down to the depth of 670 km, where the interface between the upper and lower mantle is located. The body waves give information on the whole mantle and on Earth’s core.

Seismic tomography provides a snapshot, an instantaneous picture today of a slowly, geologically evolving mantle. In principle, in order to use seismic tomography in a self-consistent way it is necessary to compare the tomographic images of recent seismic movements with those obtained from mathematical models of the evolution of Earth’s geological structures since the beginning of its formation.

Figure 2 provides an Equatorial cross-section of velocity anomalies through the mantle. The section through the upper mantle is the middle diagram, while the section through the lower mantle is the lower diagram. The diagrams show velocity anomalies, quantified in terms of percentage variations from a normal velocity, which corresponds to the velocity of a wave crossing a "normal" (not anomalous) region of the mantle. The velocity anomalies in the upper mantle correlate well with the Precambrian shields and mid-ocean ridges. Underneath the shields the velocity is higher (positive anomalies as high as 3% and 0.75% for S-and P-wave velocities), while in the hot regions of the mantle, corresponding to the mid-ocean ridges, the velocity is lower because of dissipative effects. This equatorial tomographic section shows a limited correlation between velocity anomalies in the lower mantle and those of the upper mantle.

The most noticeable feature in the lower mantle is the high velocity around the Pacific, which correlates with the Pacific subduction complex. Since velocity anomalies are related to elastic parameters and density, and therefore to the composition and temperature of mantle material and its fluid content, they can be considered as an instantaneous image of the thermal and compositional anomalies of the mantle. The effects of temperature and composition cannot be separated, and it could be that the seismic waves move more rapidly in the subcratonic upper mantle not only because it is cooler but also because it is compositionally different from the suboceanic mantle. In the interpretation of tomographic studies, problems related to error limitations and the resolution powers of the technique are important. The resolution in global studies, such as that portrayed in Figure 2, is limited to features with minimum horizontal and vertical dimensions of 2000–3000 km and 150–400 km respectively. More accurate resolution can be obtained today, particularly in regional tomographic studies.

Two main areas of low velocity occur in the mantle, in the central-south Pacific, and underneath southern Africa, providing a sort of bi-polarity. They correspond to the areas of positive anomalies of the geoid. They do not coincide with significant plate boundaries and occur beneath both oceanic and continental lithosphere, showing no apparent relationship with surface tectonics.

Figure 2. Equatorial cross-sections of seismic wave velocity anomalies, for shear waves (in the upper mantle, middle panel) and compressional waves (in the lower mantle, bottom panel). Source: Woodhouse and Dziewonski, 1984.

4. Changes in Earth’s Gravity Field and their Implications for Mantle Dynamics   

Satellite geodesy is today undergoing rapid development, because of new technologies such as laser tracking of satellites and gradiometry (measurements of gradients of the gravity field), which make it possible to detect the gravity field, and in particular its time variations, with high accuracy and spatial resolution. A constellation of geodetic satellites, orbiting around the Earth at different altitudes, is sensitive to time variations in the gravity field, which cause perturbations in their orbits. These perturbations can be monitored by means of laser tracking from laser stations located at different sites on the surface of Earth. Laser beams are shot from a station to the satellite, and reflectors mounted on the satellite send back the beam to the station: from the fly time of the laser beam it is possible to measure accurately the position of the satellite with respect to the station, and thus with respect to the Earth. After years of data sampling it is possible to measure the modification of the orbit of the satellite with respect to the Earth, and from the orbit perturbations the changes in the gravity field can be retrieved on the global spatial scale. Changes in the gravity field on a smaller scale are measured nowadays by means of instruments on board the satellites, which are sensitive to the local gradient of gravity.

The anomalies in the gravity field are ultimately caused by mass anomalies embedded inside the Earth, and their variation over time is related to the ongoing redistribution of mass inside and on the surface of the Earth. Thus it is clear that satellite techniques are today the most powerful tool available for gaining a deep insight into mantle dynamic processes. Seismic tomography presents an essentially static picture, but satellite geodesy has the capability to sample the ongoing changes in gravity. In order to understand the physics, the gravity data need to be compared with results from simulations of the processes thought to be involved, carried out using appropriate mathematical models. Comparison between the two sets of data makes it possible to invert the key parameters that control the geodynamic processes. The mathematical models can be fine-tuned after this data crosscheck, and in principle it should ultimately be possible to use them to predict the evolution over time of the geodynamic processes.

5. Dynamic Structure of the Mantle and Ice Loss in Antarctica and Greenland   

The time variations in the gravity field are sensitive to mass redistribution over the surface of the Earth and in its interior. Two major mechanisms are responsible for the changes in the gravitational field today: the Pleistocene deglaciation, which ended about 7000 years ago, and present-day mass instability in Antarctic and Greenland. The Earth is affected today by the first mechanism because of the viscous memory of the mantle, and by the second because the surface mass redistribution is ongoing. In order to understand how it is possible for melting of the huge ice sheets of the Pleistocene to affect the gravity field today, it is necessary to consider that placing a load on the surface of the Earth causes an immediate (elastic) subsidence underneath the load, while removing the load causes a corresponding uplift. This is followed by a time-dependent and spatially-dependent deformation, because of the viscoelastic flow in the Earth’s interior. This implies that if ice and water are redistributed over Earth’s surface, the solid Earth will react to it. Several thousand years ago the last great Ice Ages ended with the meltdown of the Laurentide and Fennoscandia ice sheet complexes, which covered the Polar regions with a maximum depth of about three kilometers of ice. Today the Earth is still rebounding in response to the disappearance of the Pleistocene ice sheets, with maximum uplift rates of about 1 cm yr–1 near the center of the Bothnic Gulf and near the south of Hudson Bay. This ongoing deformation of the lithosphere and mantle is responsible for variations in the gravity field.

Although the flow within the mantle is supposed to be extremely slow, with velocities of at most a few centimeters per year, it is extremely effective, since the geological processes responsible for the flow can remain active for thousands to millions of years. These long-standing processes explain why the morphology of our planet has been subject to profound modifications during its geological history.

Other sources of gravity changes are the present-day ice mass instabilities in Antarctica and Greenland, which are probably linked with climatic changes. The comparison between geodetic data and model results makes it possible to infer the key parameters that control the two major mechanisms mentioned above: the viscosity, characterizing the flow properties of the mantle, and the present-day ice mass imbalances, which are the largest contributors to ongoing mass redistribution.

Figure 3 shows how the rheological properties of Earth’s mantle, which characterize the ability of the mantle rocks to deform and to be transported within Earth’s interior, can be retrieved from modern geodetic observations by means of mathematical models of the Earth. This figure is based on two recent analyses of satellite geodetic data from two different research groups (shown by the red and green stripes). The horizontal stripes provide the observed time variations, or time derivative, of the quantity named J2, which represents the Equatorial bulge of the Earth in terms of its moment of inertia. J2 is plotted in units of 10-11 yr–1, and denotes the difference between the polar and the equatorial moment of inertia of Earth, divided by the moment of inertia that our planet would have if it were a perfect sphere, without equatorial bulge. J2 thus provides a measure of the Equatorial bulge and denotes a fundamental term in the gravity field. Its variation in time is a measure of how the gravity field is varying on the global spatial scale today.

The fact that the observed quantity (the time derivative of J2) is negative indicates that the Equatorial bulge of our planet is diminishing. This reduction in J2 is likely to be a result of mantle flow from the Equatorial regions of the mantle towards its Polar regions, which are essentially the response of the mantle to the aforementioned process of Pleistocene deglaciation. Material within the mantle is flowing towards the regions where the large ice sheets of the Pleistocene were located, in North America and northern Europe. During the Pleistocene, melting of ice occurred also in Antarctica. The disintegration of these large ice complexes produced a large amount of disequilibrium in the Earth, as explained above, and the flow of mantle material is compensating for the mass deficit left by ice sheet disintegration. This process is made visible by the uplift in the deglaciated regions of about 1 cm yr–1, as explained above.

In order to infer the viscosity of the mantle it is necessary to compare the J2 retrieved from satellite geodesy with the results of viscoelastic mathematical models of the Earth. The parameters of a model can be varied until the model correlates with the observational data. The values of the parameters that give a fit between the data and the model results are assumed to be representative of these parameters for the real Earth. In Figure 3, the J2 obtained by means of mathematical models is plotted as a function of the lower mantle viscosity. Three curves are provided (shown as solid, dashed, and dotted lines), corresponding to three different cases and different values of the parameters that control the time variations in the gravity field. The solid curve treats the disintegration of the Pleistocene ice sheets as the only contributor to the changes in gravity. To generate the dashed curve, the estimated effects of ice loss in Antarctica (at the maximum rate of melting today of –504 Gt yr–1, as proposed by the Intergovernmental Panel on Climate Change (IPCC)) were added to the effects of Pleistocene deglaciation. For the dotted curve, the estimated effects of ice mass loss in Greenland were added to the other effects considered. Ice loss in Greenland is taken to be –144 Gt yr–1, a figure again derived from the IPCC’s estimates. Modeling results, based on viscoelastic mathematical models of our planet, can thus be compared with the time variations in the gravity field obtained from satellite geodesy.

The most striking result from this figure is the sensitivity of the J2 to lower mantle viscosity variations, and the dominant effect of ice loss in Antarctica over that in Greenland. The time derivative of J2 shows that there are two broad intersections of the model results and the horizontal stripes depicting the observational data, thus providing two possible lower mantle viscosities. The lower viscosity solution, the first intersection on the right between the solid curve and the observational data, corresponds to a classical estimate for the mantle viscosity, close to 1021 Pa s.

The dashed and dotted curves show how ice loss in Antarctica and Greenland impacts upon the results of the Pleistocene deglaciation. With respect to the Pleistocene solid curves, the effect of melting in Antarctica is to displace the peak values (the minimum in the solid curve) in the direction of the observational data. Comparison between the dashed and dotted curves shows that ice loss in Greenland reinforces the effects of Antarctica on the time derivative of J2.

These results suggest that some of the scenarios proposed by the IPCC could modify our estimate of the lower mantle viscosity. It can be seen that the intersection between the modeled results and the observational data is very sensitive to estimates of present-day mass balance in the polar regions of the Earth (that is, to estimates of ice loss). It is clear that, on the basis of the results shown in Figure 3, there is a trade-off between estimates of lower mantle viscosity and the amount of present-day ice loss in Antarctica and Greenland. In fact when the effects of melting in Antarctica are added to the effects of Pleistocene deglaciation, the intersection between the model results and observations moves towards higher estimates of viscosity than the value of 1021 Pa s obtained from the old postglacial rebound models. The intersection between the dashed curve (corresponding to ice loss in Antarctica) and the data occurs for lower mantle viscosity of 2 x 1022 Pa s (Figure 3). This finding has profound implications for the style of subduction, as will be shown in Section 6. These results suggest that Earth’s mantle is strongly stratified in viscosity, with the lower mantle being more viscous than the upper mantle. Satellite geodesy and mathematical modeling of the Earth can be used to constrain not only the viscosity, but also the amount of present-day ice mass loss in Antarctica and Greenland, which is relevant for climatological modeling.

Figure 3. J2 as a function of lower mantle viscosity, for fixed viscosity in the upper mantle. The green and red stripes stand for the two satellite solutions quoted in the bibliography. The solid curve corresponds to Pleistocene deglaciation; the dashed one corresponds to Pleistocene plus maximum ice loss in Antarctica, while the dotted one includes ice loss in Greenland.

6. How the Rheology of the Mantle Impacts the Style of Subduction   

In Section 5 it was shown how it is possible to retrieve the rheological structure of Earth’s mantle. In the following, some results obtained from mathematical models of the lithosphere and mantle show the effects of a viscosity increase in the lower mantle on the style of subduction.

The results of Figure 4 are based on simulations of a subduction process in which a cold oceanic lithosphere is forced from the right to penetrate the viscous mantle at a constant velocity. The rectangular box represents Earth’s mantle, and the vertical and horizontal scales indicate the depth in the mantle (vertical scale from 0.0 to 1.0) and the distance from the left edge of the oceanic lithosphere (horizontal scale from 0.0 to 2.0), with the convention that from the bottom to the top of the box, the depth scale spans the whole mantle. The mathematical model used in these calculations is able to solve for temperature and velocity distribution in a viscous medium that simulates the Earth’s mantle. The temperature scale varies from 400 to 1700 K, and the black arrows indicate the velocity of the mantle particles; the temperature and the velocity field are obtained by solving the momentum and temperature equations by means of numerical techniques in the domain containing the viscous material in the rectangular box. The momentum equation corresponds to the fundamental law of dynamics, which states that force equals mass times acceleration of a body, translated into a form appropriate for the mechanics of the continuous media. The temperature equation represents the principle of energy conservation.

The results shown in Figure 4 are two-dimensional models of subduction, so these subduction patterns represent simulations of vertical cross-sections of real geological structures. The dark regions in the panels represent the cold lithosphere, at different times after the initiation of subduction.

In order to initiate subduction, the oceanic lithosphere at the surface is weakened in a region in the center of each panel, to create a plate boundary and to facilitate the initiation of the subduction process. This weakening is obtained by means of a reduction in viscosity, and the model allows for the possibility of varying the angle of inclination with respect to the surface of this thin channel of reduced viscosity. In the calculations portrayed in these figures, the angle has been fixed at 45º.

The top group of four panels (Figure 4a) assumes a uniform mantle, and the bottom group (Figure 4b) a stratified mantle, in which the lower mantle viscosity is higher than in the upper mantle, in agreement with the results outlined in the previous section. In Figure 4a, the four panels from top to down and from left to right, correspond to 9, 15, 22, and 33 Ma after the initiation of subduction. In Figure 4b the time intervals are the same except in the fourth panel, which corresponds to 52 Ma after the initiation of the process.

Comparison between the panels shows that a layered-type viscosity structure has a profound influence on the style of subduction. A viscosity increase in the lower mantle has the effect of inhibiting the penetration of the subducted lithosphere in the mantle and of modifying its shape, causing advective thickening of the deep portion of the subducted slab. In nature, the seismicity of slabs does in fact appear to terminate at the base of the upper mantle, and there is still controversy over whether it is possible for slabs to penetrate the lower mantle. This effect becomes even more important if subduction is not driven by active convergence (as was assumed for Figure 4a–b, where a velocity is imposed at the right edge of the oceanic lithosphere), but is driven solely by slab pull. Slab pull means that subduction is passive and is driven solely by the gravitational instability of the cold, and thus dense, part of the subducted lithosphere. This oversimplifies the composition of the mantle, which is likely to also be stratified from the compositional point of view, thus providing larger densities that move deeper than has been imaged so far, and decreasing the slab pull effect.

These results clearly show how the issues touched on in the three sections above are strongly linked. The viscosity values inferred from the comparison of dynamic models of the mantle and satellite gravity data impact the dynamics of subduction, in terms of depth penetration of the slabs in the mantle and shape of the subducted lithosphere. This in turn affects the interpretation of the seismic tomographic images described in Section 1, and plate kinematics at the Earth’s surface.

Figure 4. Numerical simulation of the subduction process, for different times after the initiation of subduction. The top panel represents a uniform viscosity mantle. In the bottom panel the lower mantle viscosity has been increased by a factor of 30 with respect to the upper mantle. Details on the numerical approach required to obtain these results can be found in Marotta and Sabadini, 1995.

7. Plate Kinematics   

The present movement of lithospheric plates is manifested and detected by seismicity and geodetic measurements. Past movements are recorded by the formation of orogens on top of subduction zones, which testify the convergence between plates, and by oceans, whose magnetic anomaly pairs make it possible to date and rate rift formation and the spreading apart of plates. Movements among plates may be considered as relative motions, but they may also be linked to independent reference frames (such as hot spots or quasars), and they are then defined as absolute motions. The movements between two plates may determine different types of tectonics: compressive, transpressive, strike-slip, transtension, and extension. A transpressive environment has compressive and strike-slip components, whereas a transtensive tectonic setting has extensional and strike-slip components.

Movements among plates may occur at any angle. Movements between plates separated by oceanic crust may be solved through magnetic anomalies: their width divided by time gives the rate of expansion. The rates measured for the Tertiary and Quaternary oceanic crust have been shown to be consistent with rates of motion detected by space geodesy; in other words, plate movements now are at similar speeds to movements in the past. Since plates move on a sphere, the relative motion between two plates can be approached with the Eulero "fixed point" theorem, where the pole of rotation of the relative motion is interpreted, and velocity among plates increases with their distance from the pole. However in nature two plates have a pole of relative motion, which is often not fixed, particularly where one plate is also rotating about an independent pole with respect to the other plate.

If we link the vectors of plate motion, which can be inferred from structural data on rift zones, transform zones and orogens of the last 47 Ma, we can plot a main stream of plate motion, as shown by the lines in Figure 5. These lines represent the average direction of motion of plates in geographic coordinates, independent from their relative motion, filtered by oblique motions (transtensive or transpressive zones) or body forces, or other factors deviating the main stress with respect to the plate vectors. For instance the direction of Arabia is inferred, considering the Red Sea as a left lateral transtensive rift and the Gulf of Aden as a right lateral transtensive rift. Only one line can represent the average main direction of Arabia moving in present geographic coordinates. It has been demonstrated that the stress field is not a direct indication of plate motion, since it may be strongly deviated at oblique plate boundaries.

Figure 5. The lines represent the direction of the main stream of plate movement, based on first order tectonic features such as subduction zones, rift zones, and oceanic transform faults. The big black arrows indicate the relative eastward mantle flow implicit with the westward drift of the lithosphere, of a few cm yr–1 detected in the hot-spot reference frame. The lines appear to be consistent for at least the past 47 Ma. Source: modified after Doglioni et al., 1999.

The flow lines generate sinusoids and not circles, so the plate motions are not described by a unique pole of rotation, but rather by a cone located near the geographic poles. A few deviations from this main trend occur. For instance the Cocos plate along the Galapagos ridge is opening at about 90° with respect to the global flow. However the length of this feature is about one or two orders of magnitude less important with respect to the general trends over Earth’s surface, and this could be interpreted as a rotation of the plate while it is moving along the flow. The flow of Figure 5 becomes unclear in the Polar Regions, where plates are also moving more slowly. The main flow lines of Figure 5 are valid for at least the last 47 Ma, from the beginning of the linear hot-spot track of the Hawaiian chain, but the Emperor-Hawaii bend in the hotspot migration has been demonstrated not to correspond to a change in plate motion of the Pacific plate. Therefore the flow lines could be valid even earlier. For South America, the Atlantic Ocean, Africa, India, and Eurasia, the mainstream flow seems to have been stable at least for the past 250 million years, based on rifts, thrust belts, and transform fault trends. The flow describes a gradual undulation from eastern Africa to Asia, where it rapidly turns at about 90° toward the Pacific. The area of this bending runs from the Barents Sea across Russia, China, Indonesia, and the north of Australia (Figure 5 ). It is interesting to note that the main continental lithosphere is concentrated along this bending.

Note that the flow lines depicting plate motions in Figure 5 do not imply the westward drift of the lithosphere, which has been detected in the hot-spot reference frame, or in the plate motions relative to Antarctica. They simply link together the main directions of movement.

At present, measurements of plate motions are made by space geodesy such as the Global Positioning System (GPS), satellite laser ranging (SLR-Lageos) and very long baseline interferometry (VLBI). Figure 6 shows the data collection of NASA, with space geodesy in the international terrestrial reference frame (ITRF) referred to a hypothetical fixed Earth’s center. Those vectors confirm the presence of a quite stable direction of motion of plates in an absolute independent reference frame: from the south and north Atlantic, plates tend to deviate toward the northeast throughout Africa, Eurasia, and the Indian and Australian plates, then plates turn to a Pacific west-northwest direction. This trend strictly overlaps the predicted directions of tectonic features shown as Figure 5. In the NASA geodetic data, for sake of simplicity, no net rotation (NNR) between the lithosphere and the interior of the Earth is imposed; therefore the westward drift is simply omitted. However, summing all the absolute vectors of plates, there remains a residual component toward the west. This result is because the Pacific plate is the fastest and biggest plate moving toward the west-northwest, and the other plates moving along opposite directions are unable to compensate for it. Plates move faster in the Equatorial and tropical areas than in the Polar Regions, suggesting that the movement is controlled in some way by Earth’s rotation, at least in terms of direction.

A basic target in tectonics at all scales is the determination of the depth of the decollements. One assumption that may be inferred for plate tectonics is that the lithosphere is detached with respect to the asthenospheric mantle. This is indicated for instance by the Hawaiian chain, which shows that the lithosphere is moving relative to the mantle. This decollement may be distributed in the asthenosphere (1017–1019 Pa s), where a strong viscosity contrast exists relative to the lithosphere (1021–1025 Pa s), and it may be variable from point to point, determining differential velocities among plates.

Figure 6. Present-day plate motions vectors inferred from space geodesy, with reference to a hypothetical fixed Earth center (ITRF, international terrestrial reference frame). Heflin et al., 2002, NASA database, , Note how geodesy largely confirms the undulating main stream of the previous figure.

8. The Hot-Spot Reference Frame and the Westward Drift of the Lithosphere   

There are a number of volcanoes that describe linear rejuvenating tracks on Earth’s surface. Some of them also erupt large amounts of magma. The best example is the Hawaiian chain, comprising several volcanoes, both subaerial and submarine, which are active in the eastern tip (for example, Mauna Loa), and are generally older to the west-northwest of the chain, up to more than 47 Ma at the farthest tip, 4000 km away (Figure 1). They persist in a given area, and their movementindicates that the lithosphere is moving on top of a deeper source. Some hot-spot tracks are located in internal zones of the plates, but other so-called hot spots are positioned on plate boundaries. The best example of the first type is again the Hawaiian chain, whereas examples of the second type are for example Iceland, Azores, Galapagos, Eastern Island, Reunion, Ascencion and Tristan da Cunha). The Hawaiian chain is independent of any plate boundary and therefore is a good reference point for anchoring measurements for plate kinematics. The Iceland chain is ambiguous for kinematic reference because it is located on the Atlantic midoceanic ridge, and the oceanic ridges are moving with respect to both each other and the underlying mantle. Some volcanic spots have been interpreted as wet spots rather hot spots: the larger water content in a cooler area of the mantle decreases the melting point, and generates large volumes of magmas at a shallow depth. Therefore there are different types of volcanic tracks, which should be separated when using the hot spots as a fixed reference frame to measure plate kinematics.

On the basis of the hot-spot reference frame, an average westward drift of the lithosphere relative to the asthenospheric mantle has been measured. When the vectors of plate motions in the hot-spot reference frame are summed, a westward component of the lithospheric motion larger than 50 mm yr–1 remains. This westward drift implies that plates have a general sense of motion, and are not moving randomly. If we accept this observation, we can conclude that plates are moving along the flow lines of Figure 5, but with different velocities, toward the west relative to the mantle. Rather than exactly west, it would be better to say that they are moving generally westward (to the southwest, the west-northwest, and so on) along the sinusoidal flow lines, which undulate and are not east–west parallel. Therefore plates are more or less detached with respect to the mantle, as a function of the decoupling at their base. The degree of decoupling is controlled by lithospheric thickness and composition, and/or the thickness and viscosity of the underlying asthenosphere, and the lateral variations of these parameters. When a plate moves faster toward west than an adjacent plate to the east, the resulting plate margin is extensional; when a plate moves faster to the west than an adjacent plate to the west, their common margin will be of the convergent type. The westward drift is measured as an average delay of the lithosphere with respect to the hot spots, but it implies a relative eastward motion of the mantle relative to the lithosphere. Lateral heterogeneities in the asthenospheric viscosity are a possible mechanism to justify this phenomenon, although its real causes are not yet fully understood. (It might for example be related to Earth’s rotation.)

9. Plate Boundaries   

9.1. Subduction Zones

Subduction is the term used for a part of the lithosphere plunging and falling down into the mantle. It occurs where plates converge, at rates that vary from a few mm yr–1 to more than 100 mm yr–1, but it also forms in the absence of convergence along some west-directed subduction zones. The hinge of the subduction moves toward the foreland, and this is called "subduction rollback" or "slab retreat." The rollback is particularly evident for west-directed subduction zones such as the Mariannas, Barbados, and the Apennines, and it is associated with back-arc opening (Figure 6). West-directed subduction zones may be very steep and deep down to 670 km, whereas east- or northeast-directed subduction zones are generally shallower and less inclined. The westward drift of the lithosphere relative to the underlying mantle could explain those asymmetries among subduction zones (Figure 7), and the peculiar geological signatures of the orogens and accretionary prisms associated with the opposite subduction zones (Figure 8).

Figure 7. West-directed subduction zones are steeper and deeper than east-northeast or north-northeast-directed subduction zones. Note that the decollement plane (marked by a dashed red line) of the eastern plate is warped and subducted in the case of the west-directed plane, whereas it ramps towards the surface in the east-northeast directed subduction, enabling the uplift of deep-seated rocks. This asymmetry may be explained by the westward drift of the lithosphere relative to the mantle, and controls the strong differences in morphology, structure, and lithology of the related thrust belts. Source: after Doglioni et al., 1999.

Figure 8. Main structural differences among orogens due to west-directed and east-northeast-directed subduction zones. The paths of three possible markers in the two systems illustrate different end-members of possible PTt metamorphic evolutions. Source: after Doglioni et al., 1999.

The westward drift supports the thesis that there is an eastward-oriented push at depth on the west-directed slabs in order to generate the typical arcuate shape of those subduction zones, like obstacles in a river, and the opening of back-arc basins. The often-overestimated convergence between Africa and Europe, as an example, has been in the order of a few millimeters per year during the Late Tertiary and Quaternary. That is one order of magnitude lower than the velocity of the Apennines subduction roll-back towards the east, while the western Mediterranean opened as a coherent system of sub-basins of the back-arc spreading in the hanging wall of the Apennines subduction (the Provençal, Valencia, Alboran, Algerian, and Tyrrhenian basins). Contemporaneously the Apennines arc generated and still generate compression and extension all around the arc, being the tectonic pair independent from what Africa and Europe are doing one relative to the other. The backarc extension stretched and boudinated the hangingwall lithosphere. Subduction zones are the areas where the largest amount of energy (90%) is released by plate tectonics. The reason is that rocks need much more energy to be deformed under compression than under extension (about 16 times). In fact among the 10 largest earthquakes of the last century, eight occurred along the eastern and western Pacific subduction zones, and the other two in the Indonesian and Himalayan subduction zones. The largest earthquake occurred along the Chile subduction (in 1960, magnitude 9.6). The shallow seismicity indicates that seismic coupling at convergent plate boundaries can vary enormously from one subduction to another. The depth of the decollement planes could also control the seismic coupling. In the accretionary wedge the depths of the decollements are a function of the thickness and rheology of the rocks involved. No clear distinction can be made between west-directed and east- or northeast-directed subductions, although it seems that boundaries where at least one of the plates is continental normally host larger earthquakes than those where there is only oceanic lithosphere.

The structure, geology, topography, and physical parameters strongly differ from orogens as a function of the polarity of subduction: that is, whether it follows or opposes the eastward relative mantle flow (Figure 8). Orogens and foredeeps in general are characterized by a number of parameters, such as the depth of the main basal decollement plane, which determines the volume of rocks accreted in the prism (Figure 9), the amount of subduction and its rate, the dip of the foreland monocline, the rate of convergence, the density and viscosity of the materials involved, the rates of uplift in the orogen and subsidence in the foredeep, and the spacing between thrusts (Figure 10). Tectonic erosion, the opposite of accretion, is the mechanism by which material from the upper plate is transferred to the lower plate as a result of dragging generated by subduction; in this particular case the decollement is rather transferred to the upper plate. This mechanism has been proposed for some shallow segments of the Andean subduction.

Figure 9. Volumes involved by orogens and accretionary prisms are a function of the length of subduction and the depth of the basal decollement. Here are examples of an idealized subduction of 200 km and incremental depths of the decollement. As the decollement becomes deeper, the volume of the orogen will be larger, and its structural and morphologic elevation greater. In nature, the basal decollement is more undulating and shallower for west-directed subduction zones, which in fact present lower volumes in the hanging wall.

Because of the overall westward drift of the lithosphere, we find in east-northeast-directed subduction that the basal decollement underlying the eastern plate reaches the surface and involves deep crustal rocks. With west-directed subduction we find instead that the basal decollement of the eastern plate is warped as well as subducted (Figure 7). Consequently thrust belts related to east- (or northeast-) dipping subduction show conspicuous structural and morphologic relief, involve deep crustal rocks, and are associated with shallow foredeeps. On the other hand, thrust belts related to west- (or southwest-) dipping subduction show relatively low structural and morphological relief, involve only shallow upper crustal rocks, and are associated with deep foredeeps as well as back-arc extension (Figure 8).

The different decollements in the two end members of subduction should control different PTt paths (Pressure, Temperature and time path), and therefore may generate variable metamorphic assemblages in the associated accretionary wedges and orogens. High pressure/low temperature metamorphism is more intrinsic to the frontal thrust belt of east- or northeast-directed subduction zones, whereas high temperature/low-pressure metamorphism is more typical of the hanging wall of west-directed subduction zones, where the asthenosphere replaces the slab at shallow levels (Figure 8).

Topographic and free-air gravimetric profiles across subduction zones show two distinct signatures. An average low topography (< 1250 m) and pronounced gravimetric anomalies characterize west-directed subduction zones. An average elevated topography (> 1200 m) and smoother gravimetric waves are peculiar to east- or northeast-directed subduction zones. These differences are particularly evident along the Pacific margins, but they persist also along the other subduction zones of the world, in the Atlantic, Mediterranean, Himalayan, and Indonesian regions. Therefore topography and gravimetry confirm the existence of two separate classes of subduction zones, which appear to be generally independent of the age and nature of the subducting slab.

Figure 10. Main features and physical parameters characterizing thrust belts along subduction zones

The geological characteristics of foredeeps and accretionary wedges suggest that these features are also distinguishable on the basis of the direction of the associated subduction. East- and northeast-directed subduction-related accretionary wedges show high relief and broad outcrops of metamorphic rocks. They are associated with shallow foredeeps with low subsidence rates (< 0.3 mm yr–1). In contrast, west-directed subduction-related accretionary wedges show low relief, and involve mainly sedimentary cover. The related foredeeps are deep, and have high subsidence rates (> 1 mm yr–1 = 1 km Myr–1). This differentiation is useful for both oceanic and continental subductions: for example, eastern and western Pacific subductions, or the east-directed Alpine and west-directed Apennines subductions. In a cross-section of the Alps the ratio of the area of the orogen to the area of the foredeeps is at least 2:1, whereas this ratio is 0.22:1 for the Apennines. These ratios explain why foredeeps related to east- or northeast-dipping subduction are quickly filled and bypassed by clastic supply, whereas foredeeps related to west-dipping subduction maintain a deep water environment for longer. These differences support the presence of an eastward asthenospheric counterflow relative to the westward drift of the lithosphere, which is able to enhance the pull-down in west-directed subduction zones and to sustain opposite subduction zones.

There are orogens and subduction zones that do not follow the flow proposed in Figure 5, such as the east–west trending southern Caribbean belt and the Pyrenees. These are orogens related to second-order rotations of plates. However those features have general geometries similar to the orogens associated with east-directed subduction zones. West-directed subduction zones show east-verging arcs of 1500–3000 km. They are usually younger than 50 Ma, whereas the opposed slab may be active for since more than 100 Ma. Following the Caribbean subduction examples, the west-directed subductions seem to develop along the retrobelt of former east-directed subduction zones-related orogens, where the oceanic lithosphere occurs in the foreland to the east (as it occured with the narrowing of the American continents to the east of the central America Cordillera). This could be applied to the onset of the Apennines subduction along the retrobelt of the Alpine–Betic orogen, where the Tethys oceanic crust was present. The pre-existing orogen is stretched and scattered in the back-arc basin. The backarc extension is internally punctuated by necks (sub-basins) and boudins (horsts of continental lithosphere).

Magmatism is most likely triggered by fluids that are released by slabs at depths of around 130–200 km. Magmatism is sensitive to the composition of the downgoing slab, the thermal state of the slab and surrounding mantle, the dip of the slab, the velocity of the subduction, the fluid content of the slab, and possibly the thickness and composition of the hanging wall plate and mantle. In the hanging wall of the Apennines subduction, as an example, we have volcanism fed by subducting both continental and oceanic lithospheres. As a hypothesis, since continental crust melts at lower temperatures than oceanic crust, we could expect magma to be generated at a shallower depth for continental crust than for oceanic crust; this could explain the closer location of the continental-related volcanism to the subduction hinge in the Apennines subduction zone.

9.2. Rift Zones

Rift zones are the areas where the lithosphere is split into two plates that are moving apart. The continental stage of rifts is very slow and longstanding (30–50 Ma or even more), with horizontal slip rates that may be in the order of 0.1–1 mm yr–1.

The continental stage of rift is accompanied by synrift sedimentation, which usually forms the typical three-fold succession of continental-alluvial red beds, evaporites, and carbonates. Continental margins may also be subdivided into volcanic or non-volcanic margins, as a function of the amount of associated magmatism. The Atlantic margins of Brazil and Greenland are classic volcanic margins, since the Cretaceous–Tertiary rift was accompanied by a large amount of lavas, whereas the Atlantic coasts of the United States and northwest Africa are non-volcanic margins, because rifting occurred without large contemporaneous magmatism. The differences in volcanic production may be related to water content in the mantle and/or other reasons such as geochemical variations in the mantle composition and temperature gradients.

The transition to the oceanic stage, the so-called "continental break-up," is marked by the break-up unconformity, where sediments below are tilted and cut by growth faults, and the sediments above are almost undeformed, testifying the moment when the stretching between the two diverging plates stops to deform the two conjugate continental margins but it is rather taken by the generation of the intervening new oceanic crust. The speed of rift production super-accelerates from 10 to 1000 times, to rates of spreading of 10–170 mm yr–1. Where the mantle is decompressed along rift zones because the lithosphere is moving apart, it upraises adiabatically to compensate for the mass deficit, following isostasy, and melts. Sea-floor spreading is usually differentiated into three main different types: slow (Atlantic ridge, 20 mm yr–1), intermediate (Indian ridge, 30–50 mm yr–1), and fast ridges (East Pacific rise, even more than 100 mm yr–1). Slow spreading generates a rift valley and a pronounced topography of the ridge, whereas fast spreading lacks a typical rift valley, and rather exhibits a smoother and deeper topography of the ridge. For these reasons, the top of the ocean crust in the Atlantic shows a more irregular morphology, being broken by several normal faults.

The depth of the ocean floor has been established as a function of its age: in fact, moving away from the ocean ridge, the lithosphere cools and subsides. The sea floor becomes 1000 m below the ridge in 10 Ma; then it subsides 1000 m more after further 10 Ma. The depth tends to stabilize at about 6000 m after 60 Ma. Therefore the depth z below the ridge crest is z = k square root of the age, where k is a constant of about 320. Thermal subsidence affects both the ocean and the neighboring continental margins. Tectonic subsidence along continental margins is relatively slow, up to 0.1 mm yr–1.

The topography of ocean ridges and rifts show a distinct asymmetry. The eastern sides of the East Pacific Rise, of the mid Atlantic ridge, of the NW Indian ridge are in average 100-300 m more elevated than the conjugate flank to the west. The asymmetry is maintained when bathymetry is plotted versus the square root of crustal age. A comparable topographic asymmetry occurs in the Red Sea and Baikal rifts where the "eastern" continental shoulders are more elevated. Applying the "westward" drift of the lithosphere relative to the underlying mantle, the depleted and lighter asthenosphere generated below the ocean ridge is shifting ‘eastward’ relative to the lithosphere, determining a density deficit below the eastern flank. As an application of this model, the ‘eastward’ migration of the Atlantic asthenosphere previously depleted at the ridge and sliding below the African continent, could eventually have contributed to the anomalous post-rift uplift of Africa.

With this relative "eastward" motion of the mantle relative to the lithosphere, a rift forms when a plate is moving "westward" faster than the plate to the east. The ridge moves at the sum of the velocity of the two plate divided by 2 (Fig. 11).

Figure 11. Kinematic 3-stage model of an oceanic rift where the mantle is considered fixed and two plates A and B have variable westward velocity. Separation of the two plates determines passive upwelling and partial melting of the asthenosphere. Eastward motion of the asthenosphere relative to the lithosphere causes migration of the density anomaly, resulting in a slower subsidence in the eastern flank of the oceanic ridge and in uplift of the eastern continent. Source: after Doglioni et al., 2003.

Typical rift zones such as the Atlantic formed where before there was a lithosphere thickened by orogens such as the Appalachians. The oceans are eventually closed again, forming the so-called "Wilson’s cycle," which postulates that rifts form on pre-existing mountain belts. This indicates that rift zones are mainly determined by heterogeneity in the lithosphere and their relations with the underlying asthenosphere, rather than being associated with deep mantle plumes.

There are different types of rift on Earth. Apart from the classic linear types that produce the main oceanic basins, back-arc spreading classically forms in the hanging wall of west-directed subduction zones, and is characterized by high rates of subsidence (0.6 mm yr–1). It is usually associated with the eastward rollback of the subduction. Examples are the Caribbean Sea, the western Mediterranean basin, the Pannonian basin, and the Japan Sea.

Other extensional tectonic environments develop on top of orogens because of the growth of the relief over the critical taper angle. These normal faults have likely decollement zones within the crust, whereas the classic rift zones have deeper decollement at the base of the lithosphere, at the interface with the asthenosphere.

Rift zones may be either concentrated in a few tens of kilometers (as with the East Africa rift), or they may form boudinated and stretched areas hundreds or more of kilometers wide (such as the Basin and Range province in western North America). The Basin and Range extension, rather than a backarc basin, it can be interpreted as the rift generated by the fast westward motion of the Pacific plate relative to North America since the slower Farallon intervening plate was lost in subduction under the continent.

Studies in outcropping ophiolites and in the polarization of shear waves in the mantle indicate that olivine crystals tend to parallel the direction of plate motion. This information supports the thesis that there are significant decollements in the mantle, and that it is flowing.

9.3. Transform Faults

Plate margins trending close to the relative motion between two plates are considered to be transform faults. These margins are also detached at the lithosphere base and generally considered as conservative margins since not significant new lithosphere is created (rift zones) or destroyed (subduction zones) They may be both oceanic (for instance, the Romanche fracture zone in the Atlantic ocean, bordering the African and South American plates) and continental (for instance, the Dead Sea fault, bordering the Arabian and African plates). Transform faults are among the longest tectonic features on Earth. Because of thermal differences, 2–4 km large vertical offsets may form along oceanic transform faults where oceanic lithospheres of different ages are juxtaposed. Along those fault scarps, complete sections of the oceanic crust with the basal Moho and the underlying mantle have been described. Transform faults are inherited from both the irregular break-up of the continental lithosphere and the intrinsic evolution of the oceanic plate margins. In terms of energy, transform faults are rather passive features, and apparently do not contribute positively to the energy budget of plate tectonics.

The San Andreas system in California has been frequently and erroneously used as a case history of a transform fault. In fact the San Andreas is a quite unique geodynamic setting, where the North American plate interacts with the Pacific plate only along the transfer zone of the Juan de Fuca ridge, with the East Pacific rise to the southeast. The Pacific–North America plate boundary along the San Andreas Fault system is notoriously a right-lateral transpressive margin where both almost pure thrust and strike-slip tectonics take place. The Pacific plate travels west-northwest, forming an angle of about 25° with the boundary. Since the Pacific is moving toward the west-northwest faster than North America, right lateral transtension should result along the San Andreas system. However the velocity of the transfer zone is slower than that of the plate. North America in turn travels westward obliquely to the boundary faster than the transfer zone, and North America overrides the same margin with a left-lateral transpressive component. Therefore the right-lateral transpression of the San Andreas system can be partitioned into a) a sinistral transpression along the southwestern margin of the North America plate, obliquely overriding b) a faster right lateral transtension occurring along the transfer margin of the Pacific plate between the East Pacific rise in the California Gulf and the Juan de Fuca ridge to the northwest. This is because of the oblique trend of the Pacific and North America plate margins with respect to their motion in an absolute reference frame.

Therefore the geodynamics of California is marked by a very unique setting in which there is a special subduction where, in contrast with classic subduction zones, the footwall of the subduction plane is obliquely diverging from the hanging wall in an east–west section, while it is converging at slower rates in a northeast–southwest direction. The extensional east–west component is absorbed into the Basin and Range rifting, whereas the compressive northeast–southwest component is mainly expressed in the Coast Ranges and California offshore. The compression perpendicular to the San Andreas is then not intrinsic in the strike-slip movement, but it is rather an independent tectonic factor. The kinematics of the San Andreas system shows how two independent tectonic settings can co-exist in an area.

Space geodesy has confirmed how the relative motion along North America and Pacific plate boundary is diffused in a number of faults covering an area a few hundred kilometers wide. Plate boundaries are in fact wide areas of deformation.

10. Plate Kinematics versus Mantle Dynamics   

The main well-established forces acting on the lithosphere and possibly controlling plate tectonics are the slab pull and eastward mantle flow. Other lower forces are the ridge push and the trench suction. The slab pull is the positive load of the slab with respect to the underlying mantle because it is cooler, assuming both are of similar composition. The energetic contribution of the Earth’s rotation is still controversial, in spite of the westward drift of the lithosphere. Motions in the mantle and of plates disturb the rotation and determine the true polar wander of the Earth’s rotation axis.

Oceanic lithosphere more easily subducts since it is denser and heavier than the continental lithosphere. There is clear evidence for this in the age of the ocean crust, which spans from 180 to 0 Ma (the oldest ages are in the western Pacific and in the margins of the central Atlantic). The continental crust dates back more than 4000 Ma, indicating its positive buoyancy and its permanence at the surface in comparison with the heavier oceanic crust, which has rather been continuously recycled. During subduction, the eclogitization of the oceanic rocks determines a further increase in density. However the lithosphere, like the whole Earth, is chemically differentiated, with lighter elements at shallow levels and heavier elements in the deepest core of the planet. Therefore a mineralogical difference may also be expected between the asthenospheric and the lithospheric mantle above it. The slab pull effect may be lower when these differences are introduced; in fact along several subduction zones the seismicity shows that the slab is under compression along its length, indicating resistance to subducting, whereas there are areas showing a slab under extension, particularly at shallow levels, indicating that the slab is really pulled down. There is unambiguous evidence that the continental lithosphere is also at least partially subducted. In order to bring light crust into subduction, the slab pull alone is insufficient and the eastward mantle flow may energetically contribute to this event.

Mantle convection is expected, because Earth is cooling and because material is uprising along oceanic ridges and downgoing along subduction zones. Moreover, lateral variations and gradients in temperature, density, fluid content, and viscosity should determine slow creeps within the mantle. However we do not know the constraints on the velocity of these movements, and none of the proposed models of mantle convection can account for the simpler pattern in plate motion we observe at the surface, nor has a unique solution been proposed for how material in the mantle convects. At the moment there is no way to link mantle dynamics and plate kinematics at the surface, considering that the mantle and lithosphere are detached. Plates appear to follow a main stream, both now and in the geologic past, whereas mantle convection is expected to generate cells with a typical rather circular-polygonalshape. Earth’s rotation is also able to generate a possible polarity in the kinematics of the cores, mantle, and lithosphere, a sort of railway path. Plates may be more detached on a relatively less viscous mantle than on a relatively more viscous mantle, therefore lateral heterogeneity in the asthenospheric mantle may determine different decoupling from the overlying lithosphere. Variations in decoupling are in turn responsible for differential velocity among plates and plate tectonics. The Atlantic and Indian ridges are in fact moving apart with respect to Africa, proving not to be fixed both relative to each other and relative to any fixed point in the mantle. This evidence confirms that ocean ridges are decoupled from the underlying mantle. Mantle convection models show the upraise and sinking of the mantle with fixed cells, with steady vertical plumes and polygonal shapes in an horizontal view; plate tectonics rather show linear features at the surface, and plate boundaries moving one respect to the other, and unstable. These tectonics are erroneously linked to horizontally moving uprising plumes and subduction zones, which are not predicted by physical convection models. In other words, mantle convection alone seems not able to generate plate tectonics. A more robust contribution of the Earth’s rotation in combination with mantle convection could be envisaged.


Thanks to Angelo Peccerillo for involving us with this article in the preparation of the EOLSS.

Related Chapters   

Click Here To View The Related Chapters


Boudins: Segments of lithosphere (or crust, or even beds) which have been stretched and isolated in thicker (boudins) or thinner elements (necks).

Decollements: Any surface where a shear has occurred between the hangingwall and the footwall, with a dip close to the bedding. It is generally assumed close to the horizontal.

Foredeeps: Basins that form at the front of the orogens either for the load exerted by the orogen itself (e.g., Alps), or for the slab retreat imposed by the "eastward" mantle flow (e.g., the Apennines); the material eroded from the orogens is accumulated in the foredeeps.

Horsts: They are the remaining higher blocks in the footwall of normal faults in extensional settings, whereas the adjacent grabens are in the hangingwall of the normal faults, being the most depressed parts.

J2: Dynamical form factor of the Earth, or ellipticity coefficient, describing the oblate shape, defined by (C–A)/MR2, where C and A denote the polar and equatorial moments of inertia of Earth, M and R denote the mass of Earth and its radius of 6371 km. It represents the term in the gravity field associated with the equatorial bulge of the Earth.Lithosphere: The cold outermost layer of Earth. Its average thickness is about 100–150 km.

Mantle: The viscous part of Earth between the lithosphere and the core, from a depth of 100–150 km down to the core–mantle boundary, at a depth of 2990 km from the surface. It is characterized by a transition zone at a depth of 670 km separating the upper and lower mantle.

Mathematical models: Simulate the behavior of Earth during the geodynamic process under study by means of mathematical algorithms.

Ophiolites: Remnants of oceanic crust accreted and sandwiched in orogens.

Orogens: Areas where material is accreted and thickened above subduction zones (e.g., Andes, Alps, Himalayas).Rheology: The study of how the lithospheric and mantle rock deforms and creeps under the application of shear stresses.

Satellite geodesy: The branch of science that deals with analysis of the field of gravity, using satellites orbiting Earth.

Slab pull: The downward force exerted by a slab because of its supposed higher density than the mantle in its cooler state.Subduction: The process of penetration of the lithosphere in the mantle.

Westward drift of the lithosphere: The average delay of the lithosphere relative to the mantle, as detected in the hot-spot reference frame.


Anderson D.L. (1989). Theory of the Earth. Blackwell, Boston366 pp. [This book provides extensive data on the composition of the Earth.]

Bally A.W. (1983). Seismic expression of structural styles. A picture and work atlas, American Association of Petroleum Geologists, Studies in Geology, Tusla, Oklahoma. v. 15, 3 vols. [This atlas contains a number of seismic examples on the structure of the upper crust.]

Cheng M.K., Shum C.K., and Tapley B.D. (1997). Determination of long-term changes in the Earth’s gravity field from satellite laser ranging observations, Journal of Geophysical Research, v. 102, B10, pp. 22,377-22,390. [This presents one of the two sets of observational data retrieved from satellite geodesy and used in Figure 3.]

Devoti R., Luceri V., Sciarretta C., Bianco G., Di Donato G., Vermeersen L.L.A., and Sabadini R. (2001). The SLR secular gravity variations and their impact on the inference of mantle rheology and lithospheric thickness. Geophysical Research Letters, v. 28. 5, pp. 855-858. [This presents one of the two sets of observational data retrieved from satellite geodesy and used in Figure 3; it also presents the methodology to retrieve the viscosity of the mantle from satellite data.]

Doglioni C., Carminati E. & Bonatti E. (2003): Rift asymmetry and continental uplift. Tectonics, v. 22, 3, 1024, doi:10.1029/2002TC001459. [This paper describes the asymmetry between the two flanks of the rift zones, possibly generated by the westward drift of the lithosphere relative to the mantle.]Doglioni C., Harabaglia P., Merlini S., Mongelli F., Peccerillo A., and Piromallo C. (1999). Orogens and slabs vs. their direction of subduction. Earth Science Reviews, v. 45, pp. 167–208. [This paper describes the differences among orogens as a function of the polarity of the associated subduction zone.]

Fowler C.M.R. (1990). The Solid Earth, Cambridge University Press., Cambridge, UK, 472 pp. [This is a classic textbook on solid Earth geophysics.]

Heflin et al., 2002, NASA. [Details the GPS time series.]

Marotta A.M. and Sabadini R. (1995). The style of the Tyrrhenian subduction. Geophysical Research Letters, v. 22, pp. 747–750.

Ricard Y., Doglioni C., and Sabadini R. (1991). Differential rotation between lithosphere and mantle: a consequence of lateral viscosity variations. Journal of Geophysical Research, v.96, pp. 8407–8415. [This paper describes the westward drift of the lithosphere and proposes a physical explanation for it.]

Smith D.E., Kolenkiewicz R., Nerem R.S., Dunn P.J., Torrence M.H., Robbins J.W., Klosko S.M., Williamson R.G., and Pavlis E.C. (1994). Contemporary global horizontal crustal motion. Geophysical Journal International, v. 119, 511–520. [This article shows techniques and results in measuring present day plate motion.]

Stein C. (1995). Heat flow of the Earth. Global Earth Physics: A Handbook of Physical Constants, AGU Reference Shelf 1 (ed. T.J. Ahrens), American Geophysical Union, Washington, D.C., pp. 144–158.

Vermeersen L.L.A. and Sabadini R. (1997). A new class of stratified viscoelastic models by analytical techniques. Geophysical Journal International, v. 129, pp. 531–570. [Presents the mathematical formulation necessary to build mathematical models of the Earth that can be used to simulate the gravity anomalies induced by Pleistocenic deglaciation and present-day mass instability in Antarctica and Greenland.]

Windley B.F. (1995). The Evolving Continents. J. Wiley & Sons, Chichester, London, New York, 526 pp. [This book presents a wide number of examples on the structure of the Earth.]

Woodhouse J.H. and Dziewonski A.M. (1984). Mapping the upper mantle: three-dimensional modeling of Earth structure by inversion of seismic waveforms. Journal of Geophysical Research, v. 89, pp. 5,953–5,986. [This describes in detail the technique of seismic tomography and provides the first images retrieved for the mantle.]

Biographical Sketches   

Carlo Doglioni is full Professor of Geology at the University La Sapienza of Rome, Italy. He was formerly at the Universities of Basilicata, Bari, and Ferrara, where he did his thesis. He has visited as researcher the Universities of Basel, Oxford, and Rice University of Houston. He works mainly on the geodynamics of subduction zones and on the structural geology of the Alps, Apennines, and other areas of the Mediterranean area. He has been AAPG distinguished lecturer.

Roberto Sabadini is full Professor of Physics of the Earth at the University of Milan, Italy. He has been Editor of the Geophysical Journal International and he is past President of the Section of Solid Earth Geophysics of the European Geophysical Society and a member of Academia Europea. His research interests are geodynamics, the study of mantle viscosity from post-glacial rebound to post-seismic deformation, dynamics and rheology of the mantle and lithosphere, and their relation with Earth’s rotation instabilities.

©Copyright 2004 Eolss Publishers. All rights reserved. Encyclopedia of Life Support Systems