Surge Tectonics
CONTENTS
Chapter 1 Why a New Hypothesis?
1.2 Former & Current Concepts of Earth Dynamics
1.2.3 Mantle Convection
1.2.4 Earth Expansion
1.2.5 Vertical Tectonics
1.2.6 Zonal Rotation
1.2.7 Continental Drift, Polar Wandering
1.2.8 Seafloor Spreading and Plate Tectonics
1.2.9 Tectonostratigraphic Terraces
1.2.10 Wedge Tectonics
1.2.11 Plate Tectonics with Fixed Continents
1.2.12 Zipper Tectonics (Spiral Tectonics)
1.2.13 Viscous Flow Model
Chapter 2 Unraveling Earth History: Tectonic Dating Sets
2.1 Data Availability
2.2 New Data Acquisition
2.2.1 Submersibles and Deep-Sea Drilling
2.2.2 Sonography
2.2.3 Accurate Bathymetry
2.2.4 Seismotomography
2.2.5 Space Geology
2.2.6 Satellite Photography
2.2.7 Satellite Radar Altimetry
2.2.8 Radar Mapping of Venus
2.2.9 Other Techniques
2.3 Data Sets Unexplained in Current Tectonic Models: Foundation for a New Hypothesis
Chapter 3 Surge Tectonics
=3.2 Velocity Structure of the Earth's Outer Shells
3.2.1 Basic Framework
3.2.2 Continents Have Deep Roots
3.3 Contraction
3.3.2 Contraction Skepticism
3.3.3 Evidence For a Differentiated, Cooled Earth
3.4 Contraction as an Explanation of Earth Dynamics
3.5 Review of Surge and Related Concepts in Earth-Dynamics Theory
3.6 Geotectonic Cycle of Surge Tectonics
3.7 Pascal's Law---the Core of Tectogenesis
3.8 Evidence for the Existence of Surge Channels
3.8.1 Seismic-Reflection Data
3.8.3 Seismotomographic Data
3.8.4 Surface-Geological Data
3.8.5 Other Data
3.9 Geometry of Surge Channels
3.9.1 Surge-Channel Cross Section
3.9.2 Surge-Channel Surface Expression
3.9.3 Role of the Moho
3.9.4 Formation of Multitiered Surge Channels
3.10 Demonstration of Tangential Flow in Surge Channels
3.11 Mechanism for Eastward Surge
3.12 Classification of Surge Channels
3.12.2 Ocean-Basin Surge Channels
3.12.3 Continental-Margin Surge Channels
3.12.4 Continental Surge Channels
3.13 K Structures
3.14 Criteria for the Identification of Surge Channels
Chapter 4 Examples of Surge Channels
=4.1 Ocean-Basin Surge Channels
=4.2 Surge Channels of Continental Margins
4.3 Continental Surge Channels
4.4 Surge Channels in Zones of Transtension-Transpression
Chapter 5 The Tectonic Evolution of Southeast Asia - A Regional Application of the Surge-Tectonics Hypothesis
5.1 Surge Tectonic Framework
5.2 Surge-Tectonic History
Chapter 6 Magma Floods, Flood Basalts, and Surge Tectonics
6.2 Descriptions of Selected Continental Flood Basalt Provinces
6.3 The Use of Geochemistry in Identifying Flood Basalts
6.4 Geochemical Comparisons among Basalts Erupted in Different Tectonic Settings
6.5 Duration of Individual Basalt Floods
6.6 Flood-Basalt Provinces and Frequency in Geologic Time
6.7 Non-Basalt Flood Volcanism in Flood-Basalt Provinces
6.8 Flood Basalts or Magma Floods?
6.9 Surge-Tectonics Origin of Magma Floods
Chapter 7 Conclusions
APPENDIX
Chapter 3
SURGE TECTONICS
3.1 Introduction
Surge tectonics is a new hypothesis quite unlike previously proposed hypotheses, although many of its component parts are based on ideas long known. We believe the hypothesis provides a comprehensive and internally consistent explanation of all tectonic phenomena without the necessity of making unsupported assumptions or ad hoc explanations. We have found nothing that surge tectonics cannot explain in a simpler way than other tectonic hypotheses. Surge tectonics draws on well-known physical laws, especially those related to Newton's laws of motion and gravity. Fluid dynamics plays an important role in surge tectonics. (For more information on the laws we utilize, those mentioned in the text are defined in the Appendix; those withing more detail are referred to two standard physics textbooks by Sears et al. [1974] and Blatt [1983]. An excellent state-of-the-art fluid-dynamics text is that by Tritton [1998]).
Surge tectonics is based on the concept that the lithosphere contains a worldwide network of deformable magma chambers (surge channels) in which partial magma melt is in motion (active surge channels) or was in motion at some time in the past (inactive surge channels). These surge channels play the role of the holes in the "hole-in-the-plate" problem ("elliptical hole problem") of civil and construction engineering, industrial engineering, and rock mechanics (..., 1913-1991). The presence of surge channels means that all of the compressive stress in the lithosphere is oriented at right angles to their walls. As this compressive stress increases during a given geotectonic cycle, it eventually ruptures the channels that are deformed bilaterally into kobergens (Fig. 2.15). Kobergens were named by Meyerhoff et al. (1992b) in honor of Austrian geologist, Leopold Kober (1921-1928). Kober observed that Alpide foldbelts have been bilaterally deformed with the northern ranges vergent toward Europe and the southern toward Africa (see Fig. 12 in M-b, 1992b). Thus, bilaterally deformed foldbelts in surge-tectonics terminology are called kobergens.
Surge tectonics involves three separate but interdependent and interacting processes. The first process is the contraction or cooling of the Earth. The second is the lateral flow of fluid, or semifluid, magma through a network of interconnected magma channels in the lithosphere. We call these surge channels for reasons that will become apparent. The third process is the Earth's rotation. This process involves differential lag between the lithosphere and the strictosphere (the hard mantle beneath the asthenosphere and lower crust), and its effects--- eastward shifts---already discussed (Table 2.3). Because lithosphere compression caused by cooling is the mechanism that propels the lateral flow of fluid, or semifluid, magma through surge channels, we discuss first the velocity structure of the Earth's lithosphere and underlying layers, then the contraction hypothesis (Earth cooling), and then the effects of contraction on the Earth's outer shells.
3.2 Velocity Structure of the Earth's Outer Shells
3.2.1 BASIC FRAMEWORK
The Earth's outer shells (Fig. 3.1) consist of a "hard" lithosphere above a "soft" asthenosphere (..., 1896-1940). The interpreted seismic structure of these two shells is given in Table 3.1, which is based on Press (1966) and Iyer and Hitchcock (1989). The asthenosphere overlies another hard shell that Bucher (1956, ...) named the strictosphere. Its seismic characteristics (at least near the upper surface of the strictosphere) also are given on Table 3.1. Of these shells, the lithosphere is especially important because visible tectonic effects provide the principal clues to the origin of tectogenesis. It has not been too long, since the seismic structure of the crust and upper mantle became sufficiently well imaged to permit more than just educated guesses about it.
Before 1958, when Revelle (1958) discovered a layer of material with a velocity of 7.3 km/s on the southern part of the East Pacific Rise, the lithosphere was perceived as consisting of 6.6-km/s crust overlying directly the 8.1-km/s mantle. What Revelle (1958) discovered was a lens or high-velocity crust, or low-velocity mantle, with a P-wave velocity of 7.0-7.8 km/s separating the "normal" crust above from the "normal" crust below. Similar lenses were found almost everywhere beneath the midocean ridge system (..., 1959-1965). By 1982, a similar lens of 7.0-7.8-km/s material was found beneath most of the Earth's rifts (Figs. 2.9, 3.2-3.8). Mooney et al. (1983) suggested that such lenses are a characteristic of extensional tectonic belts.
During the 1970s, refraction shooting in the northern Appalachians discovered a similar 7.0-km/s lens under the Acadian (Devonian) foldbelt (..., 1980-1989; and Fig. 3.13). It was not long before identical lenses beneath foldbelts were identified in many parts of the world (Figs. 2.13, 2.18, 2.21, 3.9-3.14). Good images of these lenses were recorded on reflection -seismic lines used for deep continental and oceanic tectonic studies (Figs. 2.10, 2.11, 2.14, 3.15; ..., 1988). Figures 2.10 and 2.11, from the English Channel and southwestern Queensland respectively, are particularly good images of two of these lenses near the base of the continental crust.
Mooney and Braile (1989), summarizing the present state of knowledge of the structure of the Earth's crust, inserted a 7.0-7.9-km/s lower crustal layer between the 6.6-km/s sialic crust above, and the 8.1-km/s mantle below. They showed the layer to be absent in places. In general, it is present beneath cratons, platforms, foldbelts, rift systems, and wrench-fault zones. Under cratonic areas the layer is 7-15 km thick; under tectonic belts it is thicker, ranging from 10-25 km. In many places the 7.0-7.9-km/s velocity is gradational with mantle velocities (>7.9 km/s). In these places, the Moho- is not a true discontinuity but a transition zone several kilometers thick (..., 1989). Areas were found also beneath ancient platforms and shields where the 7.0-7.8-km/s layer is up to 30 km thick. Examples include the Baltic Shield (3.5; ..., 1989) and parts of the Canadian Shield (..., 1989).
The preceding suggests that the 7.0-7.8-km/s layer is distributed rather randomly, and that its thickness in a given area is a result of random processes. These conclusions are almost unavoidable if one tries to explain the distribution and thickness with any of the Earth-dynamics hypotheses published during the last century. In surge tectonics this apparent randomness of distribution and thickness is in part predictable.
The origin of the 7.0-7.8-km/s layer almost certainly is closely related to the high reflectivities of the lower crust as described and illustrated by Klemperer et al. (1986), Klemperer (1987), Goodwin and Thompson (1988), Thompson and McCarthy (1990), and others (see Figs. 2.10, 2.11, 2.14, 3.15). We emphasize the fact that a lens (or lenses) or 7.0-7.8-km/s material underlies all of the microearthquake bands that we studied, and therefore, all of the high heat-flow bands shown on Figure 2.26.
3.2.2 CONTINENTS HAVE DEEP ROOTS
An important aspect of upper mantle and crustal structure is that continental cratons have deep crustal roots (..., 1963-1996). Contrary to general belief (..., 1987a), continental roots are fixed to the strictosphere (..., 1985-1986). This conclusion is supported by large and increasing volumes of data, including neodymium and strontium studies of crustal rocks (..., 1979). The absence, or near-absence, of a low-velocity asthenosphere beneath ancient cratons led Lowman (1985, 1986) to propose an Earth-dynamics hypothesis of sea-floor spreading between fixed continents. In this hypothesis, if sea-floor spreading takes place, it is restricted to suboceanic regions. Thus, the deep roots of continents are a major obstacle to any hypothesis requiring continental movements (..., 1985-1990).
An example of deep continental roots is presented in Figure 1.1, a seismotomographic cross section of North America. The dark shading beneath the Canadian Shield shows a root extending to 400-450 km (..., 1987). Similar deep roots are seen beneath part of all of the Earth's ancient cratons. In places, however, lenses of 7.0-7.8-km/s material containing low-velocity zones (Fig. 3.5) are present (..., 1989). Such lenses containing low-velocity layers postdate the establishment of the deep cratonic roots, as we show in subsequent sections.
3.3 Contraction
3.3.1 GENERAL
A discussion of the history and concept of contraction have been presented in Chapter 1 of this book. It will not be repeated here, but the interested reader is encouraged to review that discussion.
3.3.2 Contraction Skepticism
Many workers today either doubt that contraction is taking place or fail to see why the possibility should even be considered. Bott (1971, p. 270), expressing a common opinion, wrote that because of the success of plate tectonics in producing foldbelts, contraction now "...is irrelevant to tectonic problems." Two reviewers of the paper by Meyerhoff et al. (1992b) also expressed doubts that contraction can be taking place. However, one of them, K.B. Krauskopf (pers. comm., 1990), conceded that "...too little is known about what goes on in the [Earth's] interior for any definite statement to be made." He noted that MacDonald's (1959-1965a) models could easily be as sound today as they were in 1965 because "...not much more is known today..." about the concentration of radioactive elements in the Earth's interior.
3.3.3 Evidence For a Differentiated, Cooled Earth
The evidence is straightforward. The most salient facts follow.
1. The Earth includes several concentric shells, which are explicable only if the Earth differentiated efficiently and at a much higher temperature than today.
2. The outermost of these shells may be the oceanic crust whose thickness ranges from about 4-7 km. This crust is characterized by relatively constant thickness and fairly uniform seismic properties. Both Worzel (1965) and Vogt et al. (1969) observed that if the plate-tectonic explanation of ocean-crust generation is correct, it is a truly remarkable process that produces such a uniform layer in all ocean basins regardless of the spreading rate---1.2 dm/yr or 60 cm/yr. This uniformity is explained, however, if the oceanic crust is the outermost of the Earth's concentric shells. There are other explanations, one of which is discussed later.
3. Mehnert (1969), among several, noted that the further back one looks into the geological record, the greater is the abundance of mafic rocks. This is explained if the lithosphere has been thickening through time by cooling, as Mehnert (1969) suggested.
4. Miyashiro et al. (1982), reporting on studies of the Earth's metamorphic rocks, noted that Precambrian rocks show the highest geothermal gradients and that geothermal gradients of younger rocks generally decrease to the present time.
5. A convincing evidence that huge segments of the lithosphere have been and are being engulfed by tangential compression is the existence of the previously discussed Verschluckungszonen (swallowing or engulfment zones) of Ampferer (1906) and Ampferer and Hammer (1911). In places along such zones, whole metamorphic and igneous belts that are characteristic of parts of a given foldbelt simply disappear for hundreds of kilometers along strike (e.g., Alps: ..., 1983; K...-S... foldbelt ..., 1973; New Zealand Alps ..., 1974: ...; southern California Transverse Ranges: ..., 1984). Figures 2.23-2.24 and 3.16-3.17 illustrate the characteristics of typical Verschluckungszonen. Although Mueller (1983), Humphreys et al. (1984), and other workers considered these features to be former subduction zones, this interpretation is difficult to defend because all of these zones, regardless of age, are near-vertical bodies (1) reach only the top or middle of the asthenosphere (150 to 250 km deep) and (2) do not deviate more than 10° to 25° from the vertical (..., 1983-1984).
6. The antipodal positions of the continents and ocean basins mean that Earth passed through a molten or near-molten phase (..., 1907-1968). Such antipodal relations are unlikely to be a matter of chance or coincidence (..., 1968).
7. Theory (..., 1970) and laboratory experiment (..., 1956) showed that heated spheres cool by rupture along great circles. Remnants of two such great circles (as defined by hypocenters at the base of the asthenosphere) are active today: the Circum-Pacific and Tethys-Mediterranean fold systems. The importance of Bucher's (1956) experiment to contraction theory, in which he reproduced the great circles, is little appreciated.
8. As Earth cooled, it solidified from the surface downward. Because stress states in cooled and uncooled parts are necessarily opposite one another, compression above and tension below, the two parts must be separated by a surface or zone that Davison (1887) called the level of no strain (Fig. 3.2). We, as did Wilson (1954), equate the cooled layer with the lithosphere (Fig. 3.1). The uncooled part below is what Bucher (1956) called the strictosphere. Thus, as originally proposed by Scheidegger and Wilson (1950), Davison's (1887) level of no strain must be the asthenosphere, or a zone of no strain across which the change in stress states is gradual (Fig. 3.1). Only in a cooling Earth, which approximates a closed thermal system, can an asthenosphere form.
9. Continued cooling deepens the asthenosphere and the upper surface of the strictosphere. The stresses accumulated through cooling are relieved episodically by rupture along the great-circle fractures that are the Earth's cooling cracks or the Benioff zones of current literature. Because the lithosphere is being compressed and the strictosphere subjected to tension, the mechanics of rupture should follow the Navier-Coulomb maxiumum shear-stress theory (..., 1962). Accordingly, the lithosphere Benioff zone must dip less than 45° to a tangent to the Earth's surface (in actual fact, it dips 22° to 44°, Figs. 2.36, 3.1). In contrast, the strictosphere Benioff zone must dip more than 45° to a tangent to the Earth's surface (50° to 75°, Figs. 2.36, 3.1). Benioff (1949, 1954) discovered the change in Benioff-zone dip from lithosphere to strictosphere, but Scheidegger and Wilson (1950) recognized these dips as an expression of the Navier-Coulomb maximum shear-stress theory (Figs. 2.36, 3.1). The dip values of the lithosphere and strictosphere Benioff zones confirm that the Earth is a cooling body. (Ritsema [1957, 1960], working independently, also discovered the abrupt dip changes in the dip of the Benioff zone with increasing depth.)
An important fact concerning the Benioff zone is that the two segments, one in the lithosphere and the other in the strictosphere, do not necessarily form a single, continuous zone as depicted in diagrams and cross sections (e.g., Figs. 2.36, 3.1). Benioff (1949, 1954) found that the two segments of the Benioff zone, instead of joining near the base of the asthenosphere, may be offset for distances of 100 to 200 km (Fig. 3.18). In some places such as the Lesser Antilles arc, a strictosphere Benioff zone may not even be present below the lithosphere Benioff zone (e.g., Lesser Antilles and Scotia volcanic arcs). These facts are explained in our surge- tectonics hypothesis but not by other hypotheses. In fact, all detailed modern studies of hypocenter distribution in Benioff zones show the same clear division into a shallow, gently dipping lithosphere benioff zone and a deeper, steeply dipping strictosphere Benioff zone (Fig. 2.36; ..., 1973; ..., 1977).
Ritsema's (1957, 1960) focal-mechanism studies of shallow, intermediate, and deep earthquakes showed that the Benioff-zone dip in the lithosphere is only half of its dip in the strictosphere. An even more significant discovery made by Ritsema (1957, 1960), although he attached little importance to it, was the revelation that earthquake foci above 0.03R (approximately 180 km depth) show mainly compression, whereas those below 0.03R show mainly tension. Most earthquakes above and below 0.03R, as Scheidegger (1963) also noted, have a strike-slip component. Thus Ritsema's findings lend support to Scheidegger and Wilson's (1950) interpretation of the Benioff zone as a manifestation of the Navier-Coulomb maximum shear-stress theory.
10. Computer simulations of possible Earth thermal histories (...,1959, 1963; ..., 1961; Reynolds et al., 1966), using a broad spectrum of assumed initial temperatures and chemical compositions, show that the Earth is cooling (..., 1959-1966). The fact that Earth's Benioff zones still are active earthquake-generating zones provides strong support for this conclusion. Perhaps the strongest support comes from the Basalt Volcanism Study Project (1981) report by 101 petrologists, mineralogists, and petrographers, who wrote that the repeated extrusion of basalt to the Earth's surface through its history is proof of the Earth's long history of cooling.
11. Finally, the existence of Verschluckungszonen in the lithosphere and upper mantle also constitutes evidence that the Earth is actively cooling. Verschluckungszonen are interpreted by us to be large masses of lithosphere and upper mantle that were downbuckled into the upper mantle during tectogenesis as the lithosphere readjusted its shape to fit the underlying, cooling strictosphere (Figs. 2.24, 3.16-3.17). If this interpretation is correct, the existence of Verschluckungszonen may constitute proof that the Earth has been cooling to the present day. We discuss Verschluckungszonen in a subsequent section.
3.4 Contraction as an Explanation of Earth Dynamics
3.4.1 CONTRACTION ACTING ALONE
Despite the attraction of a cooling Earth, both Scheidegger (1963) and Bott (1971) concluded that contraction acting alone is inadequate to produce the crustal shortening measured in Earth's many tectonic belts. For both geological and seismological reasons, this conclusion appears to be well-founded. They gave several reasons; three of which and one of our own are crucial.
1. The total amount of shortening measured across the Earth's foldbelts far exceeds what can be inferred on theoretical grounds, whether one uses the contraction model of MacDonald (1963) or of Jeffreys (1970). Even if one accepts Bucher's (1955...; 1956...) outstanding demonstration that apparent (measured) shortening can be and generally is four to five times true shortening, the contraction hypothesis cannot explain all true shortening in foldbelts. (Lyttelton's [1982] theoretical estimate of 2,000 km of shortening adequately explained the measured shortening, but his hypothesis requires cataclysmic geological events that need to be sought in the field.)
2. Contraction alone is unable to explain the origins of all types of tectonic belts---compressional foldbelts, tensional rift zones (including midocean ridges), and strike-slip zones.
3. Ritsema (1957, 1960) and Scheidegger (1963) observed that earthquake first- motion studies show that strike-slip motions are most common in Benioff zones, not just in strike-slip and rift zones. Contraction alone cannot explain the ubiquitous strike-slip component.
4. Contraction theory requires that foldbelts are concentrated in and adjacent to oceanic trenches. This is not observed. More than 50% of the Earth's foldbelts lie at great distances from the surface trace of a Benioff zone, and all Jurassic- Cenozoic foldbelts lie within the high heat-flow bands illustrated on Figure 2.26. This cannot be explained by any Earth-dynamics hypothesis yet proposed. However, if contraction could lead to tectogenesis of large parts of the lithosphere far removed from Benioff zones, the preceding objections to the contraction hypothesis would be irrelevant.
3.4.2 CONTRACTION ACTING AS THE TRIGGER FOR TECTOGENESIS
Figure 2.26, as we have discussed, portrays the bands of high heat flow that crisscross the Earth's surface. We believe that this reticulate network of high heat-flow bands is underlain by a network of interconnected magma chambers. In our surge-tectonics hypothesis, these magma chambers are the mantle diapirs discussed in preceding sections and summarized in part in Table 2.1. Figure 2.26 indicates that these mantle diapirs, or magma chambers, are interconnected. The interconnected channels comprise the surge channels of surge tectonics. If the hot material in these channels is sufficiently mobile, lateral flow through them should be possible provided a pressure gradient is present. The compression already present in the lithosphere would provide the force needed to initiate and maintain flow. We emphasize that such flow would be temporally discontinuous (.i.e., episodic) and, when it did occur, would be extremely slow. Flow velocities are discussed later.
We have pointed out that in a cooling Earth, the lithosphere by definition is everywhere and at all times in a state of compression (Fig. 3.1; ..., 1887-1982). The compression is concentrated in planes tangent to the Earth's surface, and is equal in all directions. James C. Meyerhoff (pers. comm., 1988) called this equiplanar tangential compression, and this compression in the lithosphere is what Bucher (1956) referred to (incorrectly) as all-sided compression.
The only elements in the lithosphere that disturb this approximately equiplanar tangential stress state are the surge channels. Flow can take place in these channels wherever a pressure gradient develops. For example, the escape of lava from a channel lowers the pressure at that point, and equiplanar compression, acting at right angles to the surge-channel walls, mobilizes the fluid elements inside the channel until pressure equilibrium is restored. The presence above active channels of channel-parallel fault, fracture, and fissure systems indicates that (1) flow takes place along the full length of each tectonic belt and (2) the channels are in communication with the Earth's surface through the fault-fracture- fissure system.
Surge channels and their fault-fracture-fissure systems constitute zones of weakness in the lithosphere. Because (1) the channels are the only bodies in the lithosphere that, owing to their potential to contain mobile fluids, they have the capacity to upset the state of equiplanar tangential compression, and because (2) they are constantly losing their contents to the surface lithosphere compression ultimately destroys them. Their deformation and ultimate destruction are the essence of tectogenesis. Thus the cooling process in the Earth's strictosphere effectively guarantees the presence within the lithosphere of a powerful mechanism for tectogenesis.
3.5 Review of Surge and Related Concepts in Earth-Dynamics Theory
Several workers proposed, on both theoretical and geological-geophysical grounds, the presence of bodies similar to surge channels in the lithosphere. Others developed concepts much like those that are the basis for surge tectonics.
3.5.1 SURGE CHANNELS
Our study of this topic was by no means exhaustive, we may have missed important references that deal with the concept of surge-channel-like edifices in the lithosphere and uppermost mantle. A particularly good example of a surge-channel- like feature was proposed by Vogt (1974) for the Iceland region, including the Kolbeindey, Reykjanes, and Faeroe-Greenland ridges. He wrote (...), "In the model I assume there is a pipe-like region below the spreading axis, extending subhorizontally away from a plume such as Iceland.... This mid-oceanic pipe extends from the base of the axial lithosphere, about 5 or 10 km deep, down to maximum depths (30 to 50 km?) from which basalt melts segregate and rise. Tholeiitic fluids would be released from the entire pipe; origin depths of 23 km for the Mid-Atlantic Ridge and 16 km for the East Pacific Rise ... Would approximate depth to the center ot the pipe. The ultrabasic mush in this pipe is assumed to be flowing away from the hot spot at a rate determined by pipe diameter, viscosity, and horizontal pressure gradient...." Vogt (1974) estimated that flow ranged laterally outward from Iceland from 500 to 600 km. A similar study by Gorshkov and Lukashevich (1989) suggested that channels are present beneath the full length of the midocean ridge system. According to them, flow beneath the Mid-Atlantic Ridge would be from hot spots located beneath Antarctica in the south and Iceland in the north, with the two flows converging near the equator.
Other midocean ridges where some type of axis-parallel flow and/or rift propagation have been postulated include the Juan de Fuca Ridge (..., 1975) and the Galapagos Rift (..., 1977). After the discovery of systematic segmentation along the midocean ridges beginning with the East Pacific Rise (..., 1982), Lonsdale, MacDonald, Fos, and others commenced a series of investigations that led to a general postulate of ridge-parallel flow in the midocean ridges (..., 1988-1989). Macdonald et al. (1988) proposed that mantle diapirs rise beneath the centers of each ridge segment, thereby accounting for the greater elevations of the central parts of such segments. From the crest, ridge-parallel lateral flow commences. Such flow halts in the depressed areas between adjacent segments because of mutual impingement.
Sonographs of the midocean ridges show that the Macdonald et al. (1988) hypothesis is not tenable. Where the diapirs rise in the centers of the ridge segments, radial and/or annular structures should be present. Where the flows from adjacent segments impinge, compressional structures should be present. Neither structural form is observed. Instead, linear fractures and faults extend for hundreds of kilometers along strike, with interruptions only at transform faults, we traced individual fault traces through the fault zone from one ridge segment to the next, which negates the Macdonald et al. (1988) hypothesis.
Surge-channel, or interconnected mantle-diapir systems have been reported from small oceanic basins and continental areas. A well-known example of the former is the Tyrrhenian Sea west of Italy, where geophysical techniques and very high heat flow show the presence of a very large diapir-like body at shallow depths (..., 1988).
Among continental examples, the Fergana Valley in Soviet Central Asia is the best documented Kuchay and Yeryemin (1990) discovered a very large pipelike body, or channel, below this large east-west-striking late Cenozoic structural depression nestled among the western ranges of the Tian Shan. In their summary (p. 45), Kuchay and Yeryemin concluded: "Interpretation and interpolation of geophysical data from the [Fergana Valley] lead to the conclusion that the base of the 'granitic' layer undergoes partial melting, as a consequence of which there forms a layer of lowered viscosity that has been identified seismically as a zone of reduced velocity above the Conrad discontinuity. It is possible to consider this zone as a 'granitic' asthenochannel, a subhorizontal layer that has a high strain rate and in which relative lateral displacement takes place between the contents of the asthenochannel and the surrounding rock layers. The absence of a gravity anomaly, which is an indication of variations in thickness and other properties within the zone of reduced velocity, suggests that the 'granitic' layer and the 'granitic' asthenochannel have the same density."
3.5.2 USE OF THE SURGE CONCEPT IN TECTONICS
We have found three parallel and presumably independent derivations of the surge- tectonics concept. The most recent of these originated with Hollister and Crawford (1986) who used "tectonic surge" to describe rapid vertical uplift in the structural core of a bilaterally deformed foldbelt (i.e., in the center of what we call a kobergen). Hollister and Crawford (1986, ...) opined that "Weakening of the crust [by] anatexis and accompanying development of melt-lubricated shear zones..." is essential to rapid vertical uplift. Paterson et al. (1989, ...) referred to this type of tectonics as surge tectonics. Tobisch and Paterson (1990) used this term in two or three areas of southeastern Australia. Their usage is close to our own.
The third and, as far as we can determine, oldest use of the term in tectonics in recent literature was by Meyerhoff and Meyerhoff (1977). They proposed that asthenosphere surges (1) from beneath the Asian continent, (2) between North and South America, and (3) between South America and Antarctica produced the eastward- facing island arcs in the three regions. The idea was used subsequently to explain the complexities of Caribbean tectonics (..., 1990). Morris et al. (1990) used the term surge tectonics (coined by Bruce D. Martin). The paper was written during 1987-1988 and was submitted to and accepted by the Geological Society of America in 1988. Regardless, the Paterson et al. (1989) use of the term in print precedes by five months that by Meyerhoff et al. (1989). The term surge tectgonics is used in this book in the same sense that it was employed by Taner and Meyerhoff (1990).
3.6 Geotectonic Cycle of Surge Tectonics
The asthenosphere alternately expands (during times of tectonic quiescence) and contracts (during tectogenesis). Thus, when the asthenosphere is expanding, the surge channels above it, which are supplied from the asthenosphere, also are expanding; and when tectogenesis takes place, the magma in the surge channels is expelled. Tectogenesis is triggered by collapse of the lithosphere into the asthenosphere along the 30° -dipping lithosphere Benioff zones. The following is Meyerhoff et al.'s (1992b) interpretation of the approximate sequence of events during a geotectonic cycle (Fig. 3.19).
1. The strictosphere is always contracting, presumably at a steady rate, because the Earth is cooling.
2. The overlying lithosphere, because it is already cool, does not contract, but adjusts its basal circumference to the upper surface of the shrinking strictosphere by (1) large-scale thrusting along the lightosphere Benioff zones, and (2) normal- type faulting along the strictosphere Benioff zones. These two types of deformation, one compressive and the other tensile, are complementary and together constitute an example of the Navier-Coulomb maximum shear stress theory (..., 1962 -1979).
3. The large-scale thrusting of the lithosphere is not a continuous process, but occurs only when the lithosphere's underlying dynamic support fails. That support is provided mainly by the softer asthenosphere and frictional resistance along the Benioff fractures. When the weight of the lithosphere overcomes the combined resistance offered by the asthenosphere and Benioff-zone friction, lithosphere collapse ensues. Because this process cannot be perfectly cyclic, it must be episodic; hence tectogenesis is episodic.
4. During the anorogenic intervals between lithosphere collapses, the asthenosphere volume increases slowly as the strictosphere radius decreases (Fig. 3.19). The increase in asthenosphere volume is accompanied by decompression in the asthenosphere.
5. Decompression is accompanied by rising temperature, increased magma generation, and lowered viscosity in the asthenosphere, which gradually weakens during the time intervals between collapses.
6. Flow in the asthenosphere is predominantly eastward as a consequence of the Earth's rotation (Newton's Third Law of Motion; ..., 1977). Magma flow in the surge channels above the asthenosphere also tends to be eastward, although local barriers may divert flow in other directions for short distances. Coriolis force also must exert an important influence on asthenosphere and surge-channel flow, which by its nature is Poiseuille flow. Therefore, the flow at the channel walls is laminar and is accompanied by viscous, or backward drag. The viscous drag produces the swaths of faults, fractures, and fissures (streamlines) that are visible at the surface above all active tectonic belts. These bands or swaths are examples of Stoke's Law (one expression of Newton's Second Law of Motion).
7. During lithosphere collapse into the asthenosphere, the continentward (hanging wall) sides of the lithosphere Benioff zones override (obduct) the ocean floor (..., 1906-1911). The entire lithosphere buckles, fractures, and founders. Enormous compressive stresses are created in the lithosphere.
8. Both the lithosphere and the strictosphere fracture along great circles at the depth of the strictosphere's upper surface, as predicted by theory (..., 1959-1976) and demonstrated in the laboratory (..., 1956). Only two partial great circle fracture zones survive on the Earth today. These include the fairly extensive, highly active Circum-Pacific great circle and the almost defunct Tethys- Mediterranean great circle.
9. When the lithosphere collapses into the asthenosphere, the asthenosphere- derived magma in the surge channels begins to surge intensely. Wherever the volume of the magma in the channels exceeds their volumetric capacity, and then compression in the lithosphere exceeds the strength of the lithosphere that directly overlies the surge channels, the surge-channel roofs rupture along the cracks that comprise the fault-fracture-fissure system generated in the channel by Poiseuille flow before the rupture. Rupture is bivergent, whether it forms continental rifts, foldbelts, strike-slip zones, or midocean rifts. The foldbelts develop into kobergens, some of them alpinotype and some them germanotype. The tectonic style of a tectonic belt depends mainly on the thickness and strength of the lithosphere overlying it (Fig. 3.19).
10. Tectogenesis generally affects an entire tectonic belt and, in fact, may be worldwide; the worldwide early to late Eocene tectogenesis is an example (M-b, 1992b) This indicates that the lithosphere collapse that generates tectogenesis transmits stresses everywhere in a give belt at the same time. Thus Pascal's Law is at the core of tectogenesis. Sudden rupture and deformation of surge channels may therefore be likened to what happens when someone stamps a foot on a full tube of toothpaste. The speed or rapidity of tectogenesis, then, is related to the number of fractures participating in the event, as well as to the thickness of lithosphere involved, the size of the surge channel or surge-channel system, the volume and types of magma involved, and related factors.
11. Once tectogenesis is completed, another geotectonic cycle or subcycle sets in, commonly within the same tectogenic belt.
3.7 Pascal's Law---the Core of Tectogenesis
Pascal's Law (or theorem) states that pressure applied to a confined liquid at any point is transmitted undiminished through the fluid in all directions and acts upon every part of the confining vessel at right angles to its interior surfaces and equally upon equal areas. This law applies in part to all fluids, but wholly to Newtonian fluids; it is the principle behind every hydraulic machine, notably the hydraulic press. A most important condition of Pascal's Law is that the pressure (force per unit area) acts equally upon equal areas. This condition lies at the very core of tectogenesis.
The Earth, according to our surge-tectonics hypothesis, is a very large hydraulic press. Such a press consists of three essential parts: a closed vessel, the liquid in the vessel, and a ram or piston. The collapse of the lithosphere into the asthenosphere is the activating ram or piston of tectogenesis. The asthenosphere and its overlying lithospheric surge channels---which are everywhere connected with the asthenosphere by vertical conduits---are the vessels that enclose the fluid. The fluid is magma generated in the asthenosphere. The magma fills the lithosphere channels. When the piston (lithosphere collapse) suddenly compresses the channels and the underlying asthenosphere, the pressure is transmitted rapidly and essentially simultaneously through the worldwide interconnected surge-channel network; the surge channels burst and the tectogenesis is in full swing. The compression everywhere of the asthenosphere compensates for the fact that the basaltic magma of the surge channels is non-Newtonian.
A possible objection to this simple picture of tectogenesis is that the sudden application of pressure against the surge channels would consolidate the magma in the channels, and thereby prevent the bursting of the channel roof. This would be true if the channels had no communication with the surface at the onset of tectogenesis, but this is not the case. As Meyerhoff et al. (1992b) noted, the channels are connected to the surface by swaths of belt-parallel faults, fractures, and fissures.
A second possible ofjection is that the magma in surge channels is non-Newtonian; i.e., it is too viscous to transmit the added stress to all of the interconnected parts of the surge-channel system. This objection would be valid for a tectonic model in which the added stress is applied only at a single point in the system. In a contracting Earth, however, compression in the lithosphere is omnipresent. Hence, the added stress is applied everywhere along the interconnected lithosphere channels so that the viscosity argument is invalid; the added stress is being applied at an infinite number of points in the system. As shown in Figure 3.19, the thickness of the lithosphere overlying each channel is extremely important, because the thickness determines the resulting tectonic style of the channel during tectogenesis---rift valley, germanotype foldbelt, alpinotype foldbelt, midocean rift, and so forth.
Although Pascal's Law applies to all tectonic settings, it is especially important in midocean ridge systems. The law states that pressure applied to a confined liquid acts equally on equal areas of the walls of the confining vessel. The surge channels beneath midocean ridges can be thousands of kilometers wide. Hence, they are much larger than their continental-margin and continental counterparts. Morevoer, the lithosphere above midocean ridges is only 10 to 15 km thick, less than half of the thickness found in a continent-margin/continental setting. This means that the total force acting on the walls of a midocean-ridge surge channel is vastly greater than in any other setting. Thus, during tectogenesis, midocean ridges presumably rupture throughout their lengths and across widths far greater than those of continental surge channels, thereby producing veritable magma floods on the ocean floors. If one keeps in mind the fact that the most massive Phanerozoic continental flood volcanism took place from Late Permian through Middle Jurassic time (...), such magma flooding in the oceans during the same time interval would account for the fact that the oldest basalts thus far penetrated by deep sea dirlling beneath the abyssal plains are Middle Jurassic. On the midocean ridges themselves, however, basalts older than Middle Jurassic are common (Meyerhoff et al., 1992a).
3.8 Evidence for the Existence of Surge Channels
3.8.1 SEISMIC-REFLECTION DATA
As noted above, reflection-seismic techniques (...) have shown that the continental crust of the upper lithosphere is divisible in a very general way into an upper moderately reflective zone and a lower highly reflective zone (...). Closer scrutiny of the newly-acquired data soon revealed the presence in the lower crust of numerous cross-cutting and dipping events. When many of these cross-cutting events were preceived to be parts of lens-like bodies, various names sprang up: .... Strictly nongenetic names include lenses, lenticles, lozenges, and pods (...). Finlayson et al. (1989) found that the lenses have P-wave velocities of 7.0-7.8 km/s, commonly with a low-velocity zone in their middle. Thus we equate the lenses with the pods of "anomalous lower crust" and "anomalous upper mantle" that we discussed in a preceding section. Klemperer (1987) noted that many of the lenses are belts of high heat flow. Hyndman and Klemperer (1989) observed that the lenses generally have very high electrical conductivity.
Meyerhoff et al. (1992b) discovered that there are two types of undeformed reflective lenses, and that many of these lenses have been severely tectonized. The first type of lens is transparent in the middle (Fig. 3.29); the second type is reflective throughout (Fig. 2.11). Tectonized lenses also may have transparent interiors, or parts of interiors; many, however, are reflective throughout (Fig. 3.21). Where transparent zones are present (Fig. 3.20), bands of high heat flow, bands of microearthquakes, belts of high conductivity, and bands of faults, fractures, and fissures are present. Where a transparent layer is not present, high heat flow and conductivity, however, are commonly still present. Meyerhoff et al. (1992b) also found that lenses with transparent interiors are younger than those without transparent interiors; moreover, there is a complete spectrum of lenses from those with wholly transparent interiors to those without.
The best explanations of thes observations are that (1) the lenses with transparent interiors are active surge channels with a low-velocity zone sandwiched between two levels of 7.0 to 7.8 km/s material; (2) the lenses with reflective interiors are former surge channels now cooled and consisting wholly of 7.0 to 7.8 km/s material; and (3) the tectonized lenses are either active or former surge channels since converted into kobergens by tectogenesis.
3.8.2 SEISMIC-REFRACTION DATA
After Revelle (1958) discovered the presence of a body of 7.0-7.8 km/s material on midocean ridges (the East Pacific Rise), a similar body was discovered on the Mid- Atlantic Ridge, and the general lens shape was reported for the first time (...). Subsequently Talwani et al. (1965) combined gravity and seismic data, and detailed the lens shape of the surge channel across the entire Mid-Atlantic Ridge (Fig. 2.27). Fuchs and Landismann (1966) found a similar but much smaller lens beneath the Upper Rhine graben with a P-wave velocity of 7.6 km/s. Now such a 7.0-7.8-km/s lens is known to underlie every well-sutdied tectonic belt, regardless of tectonic origin. Figure 3.6 shows a 7.0-7.8-km/s lens under the northern Appalachians (...). Meyerhoff et al. (1992a, 1992b) summarized the worldwide evidence for the presence of such lenses under every type of tectonic belt. These same authors showed that the lenses under older tectonic belts contain no low-velocity zone, but that lenses in younger tectonic belts contain low-velocity zones.
3.8.3 SEISMOTOMOGRAPHIC DATA
Seismotomographic data, wherever detialed studies have been made, indicate that the lenses seen in seismic-refraction and seismic-reflection studies form an interconnected, reticulate network in the lithosphere. Although only one highly detailed seismotomographic study has been made on a continental scale---this in China, it leaves no room for doubt that the 7.0-7.8-km/s lenses with transparent interiors and the seismotomographically detected low-velocity channels in the lithosphere are one and the same (...). Figure 2.31 shows the active surge channels at a depth of about 50 km in southwestern China. Figure 3.9 is a seismotomographic cross section to a depth of 300 km across the Yunnan surge channel shown in Figure 2.31. Figure 3.14 is a more detailed cross section (above a depth of 65 km) of the central part of Figure 3.9. Velocity data from refraction surveys have been added. The 7.6-7.9 km/s layer below 50 km is the Yunnan surge channel shown in Figure 2.31; the low-velocity layer between 22 and 44 km (5.4-6.0 km/s) is a part of the Yunnan surge-channel complex (...). Using seismotomographic techniques, it will be possible to map active surge channels over the world with comparative ease. The reader should note that Figure 3.22 shows a seismotomographic image of the active kobergen of the Hengduqn Shan-Shaluli Shan that overlies the Yunnan surge channel, a further demonstration of the validity of our tectonic interpretation.
3.8.4 SURFACE-GEOLOGICAL DATA
Direct evidence for the existence of surge channels comes from tectonic belts themselves, and from one type of magma flood province. The latter include rift igneous rocks that crop out nearly continuously for their full lengths. Examples include the rhyodactic Sierra Madre Occidental-Sierra Madre del Sur extrusive and intrusive belt of Mexico and Guatemala, some 2,400 km long; the 1,600-km-long Sierra Nevada-Baja California batholith belt; the 4,000-km+ batholith and andesite belt of the Andes south of the equator; the 4,000-km-long Okhotsk-Chukotka silicic volcanic belt; the 5,800-km-long Wrangellia linear basaltic province extending from eastern Alaska to Oregon, which erupted in less than 5 Ma; and many other similar continental magma belts. The ocean basins are equally replete with them, ranging from the 60,000-km-long midocean ridge system through the 5,800-km-long Hawaiian- Emperor island and seamount chain to many similar belts of shorter lengths. Geochemical studies also show that most of these belts are interconnected. Another linear flood-basalt belt, which has been studied only relatively recently, is the subsurface Mid-Continent province that extends 2,400 km from Kansas through the Great Lakes to Ohio (Figs. 3.23, 3.24).
3.8.5 OTHER DATA
Other data mentioned in the preceding sections corroborate the interconnection of active surge channels. One of these is the coincidence of the 7.0-7.8-km/s lenses of the active surge channels (Figs. 2.9, 2.31, 3.6, 3.9, 3.14, 3.20) with the belts of high heat flow (Fig. 2.26) and with belts of microseismicity. Both the presence of high heat flow and microseismicity indicate that magma is moving within active surge channels.
However, an even more dramatic example is the June 28, 1992, Landers, California, earthquake-related activity shown on Figure 3.25. This figure shows that the 7.5- magnitude earthquake was strong enough to affect areas up to 1,250 km from the epicenter (...) and provides an exampole of Pascal's Law in action. Given the importance of Pascal's Law in surge-channel systems, the fact should be noted that the viscosity of the magma in the surge channels affected by the Landers event is sufficiently low that, when the stress was applied at a single hypocentral point (Landers), the effects could still be transmitted for 1,250 km!
3.9 Geometry of Surge Channels
3.9.1 SURGE-CHANNEL CROSS SECTION
In cross sections, surge channels have a variety of shapes, and are of many sizes and depths within the lithosphere (M-a). Two models proposed for sill-and-laccolith complexes may be ideal representation of surge-channel complexes because, despite the differences in scale, the same physical principles apply. Corry (1988) published the "Christmas Tree" model shown in Figure 2.8; Bridgwater et al. (1974) published the more complex model shown in Figure 3.26. Either of these could be cross sections of surge channels. Both are multitiered with one or more magma chambers above the main chamber. Both are formed on the basis of Newton's Law of Gravity or, more specifically, the Peach-Kohler climb force (...).
Seismotomographic images are available from hundreds of surge channels in different parts of the world. They are complex structures as Figures 3.9 and 3.27 demonstrate. Figure 3.27 is a tracing of a seismotomographic image of the multitiered Yunnan surge channel (Fig. 2.31 ...). It is also quite large for a continental channel.
Figure 3.23 shows a partly deformed ("kobergenized") inactive Proterozoic surge channel that underlies the Midcontinent Rift of central North America (...). A surge-tectonic structural interpretation is shown on Figure 3.24. The figures show that the channel before deformation consisted of a large lower chamber at the Moho-, and a smaller, higher chamber at about 20 km. Kobergenic structure developed during tectogenesis (Fig. 3.24). The P-wave velocity in the channel is 7.0 to 7.2 km/s, and the top of the channel extends to within 10 km of the surface. The Midcontinent Rift was a major flood-basalt province of 1,100 Ma.
3.9.2 SURGE-CHANNEL SURFACE EXPRESSION
Study of Figures 2.8, 2.9, 2.11, 2.31, 3.6, 3.9, 3.13, 3.14, 3.20, 3.23 and 3.24 might lead one to believe that surge channels are everywhere fairly simple structures expressed at the surface by a single belt of earthquake foci, high heat flow, bands of faults-fractures-fissures (streamlines), and related phenomena which, during tectogenesis, deform into a single kobergen. Although this simple picture is true of many kobergens, it is not true of all. Study of Figures 3.26 and 3.27 suggests that, during tectogenesis of the surge-channel complexes shown on these figures, two or more parallel kobergens may exist at the surface. Such a complex surface expression is in fact quite common. Well-documented examples are found in the Western Cordillera of North America, the Mediterranean-Tethys orogenic belt (including the Qinghai-Tibet Plateau), and the Andes, inter alia. Within the Western Cordillera, the Qinghai-Tibet Plateau, and the Andes, we have found four or more parallel kobergens side by side at the surface as documented and illustrated by Meyerhoff et al. (1992b).
3.9.3 ROLE OF THE MOHOROVIC DISCONTINUITY
The principal forces acting on the lithosphere are compression, rotation, and gravity. We have discussed the first two briefly and now endeavor to describe gravity's role in lithosphere development.
Gravity controls the depths of all magma chambers which, because the Earth cools at the asthenosphere level, have to be in the upper asthenosphere and lithosphere. The mantle above the asthenosphere is believed to be peridotite from which basalt has already been extracted (Ringwood, 1979). It has a P-wave velocity just below the Moho- greater than 7.9-8.0 km/s. The oceaenic crust above the Moho- has been assumed until very recently to have a P-wave velocity of 6.8 km/s and less; the continental crust above the discontinuity has been assumed to have a P-wave velocity of about 6.5 km/s and less.
The liquid generated by the removal of basalt from the mantle cannot be ordinary basalt, but may be akin to tholeiitic picrite (Green et al, 1979). In any case, it is less dense than the peridotite above the asthenosphere and more dense than the basalt found in deep-sea drillholes. Therefore, while in the asthenosphere, it is gravitationally unstable. Consequently Newton's Law of Gravity works to bring the magma upward to a level where it is gravitationally stable. The mechanism for this is the Peach-Kohler climb force (Weertman and Weertman, 1964; Weertman, 1971). This force compels the magma to rise through available conduits to its level of neutral buoyancy, i.e., the level where the lithosphere density and the magma density are the same (...). At this level the magma can only move horizontally.
As we have stated repeatedly, the P-wave velocity of the walls of active surge channels is in the 7.0-7.8-km/s range; that of inactive channels is 7.0-7.8 km/s throughout. In the crustal models generally accepted until the late 1980s, there was no layer having a velocity in the 7.0-7.8-km/s range. In recent years, however, a lower crustal layer has been reported in large areas of North America and elsewhere with velocities in the 7.0-7.8-km/s range (...). Mooney and Meissner (1992) in fact state that the layer is omnipresent, is ca. 3.5 km thick, and is a transition zone between the mantle and the crust.
Thus, when the postulated tholeiitic picrite magma reachs the Moho- (i.e., the zone between 8.0-km/s mantle below and 6.6-km/s above), it has reached its level of neutral buoyancy and spreads laterally. Under the proper conditions---abundant magma supply and favorable crustal structure---a surge channel can form. We suggest the possibility that the entire 7.0-7.8-km/s layer may have formed in this way. In support of this suggestion, we note that the main channel of every surge channel studied, from the Archean to the Cenozoic, is located precisely at the surface of the Moho-. This indicates that the discontinuity is very ancient, perhaps as old as the Earth itself. This fact and the great difference in P-wave velicities above and below the Moho- surface suggest in turn that the discontinuity originated during the initial cooling of the Earth. Hence, Mooney and Meissner's (1992) "transition zone" was the level of neutral buoyancy at the time the 7.0-7.8-km/s material was emplaced.
The suboceanic Moho- is some 10-15 km below sea level; its subcontinental counterpart is deeper, at about 40 km below sea level. Hence the neutral buoyancy levels of the two regimes are separated vertically by 25 to 30 km. This fact probably is closely related to the fact that surge channels underlie all continental shelf-slope breaks. Thus these breaks are transition zones between continental and oceanic regimes. MacDonald (1963, 1964) long ago stated that this transition zone must be the site of vertical faulting. Writing of the differences in the vertical distribution of radioactive heat sources under continents and ocean basins, MacDonald (1963, p.589) stated that these differences ensure "...that the boundaries between them [continents and ocean basine] become discontinuities in subsurface vertical motion. The boundary regions are thus narrow regions where faults are formed and volcanic activity is concentrated." In summary, the ocean- continent boundary is a weakness zone due to differing crustal thicknesses and compositions on either side of that boundary, and the accompanying changes in temperature regimes, specific gravities, and velocities. This weakness zone is therefore a focus for faulting and ascending magma.
It is a fact that volcanism does take place along continental mragins, not just in the "active" margins, but in the "passive" ones as well, as for example along the eastern seaboard of North America (...) and the Atlantic margin of Africa (...). In fact, a recent study of North America's eastern seaboard by Hollbrook and Kelemen (1993) shows the presence beneath the outer shelf of a linear pod of volcanic material at the Moho-. The pod has a P-wave velocity of 7.2-7.5 km/s and is at least 3,300 km long. It should be added that the above explanation does not mean that all surge channels form at ocean-continent boundaries, rather, surge channels are found in every tectonic setting, one of which is the ocean-continent boundary.
The formation of the Christmas-tree-like structures (Figs. 2.8, 3.26) at the Moho- is simply an extension of the larger scale process of magma transfer from the asthenosphere to the discontinuity. Once surge channels are established at the discontinuity, the same processes take over that brought the magma to the discontinuity in the first place, specifically, magma differentiation in the channels and the Peach-Kohler climb force (...). After lighter magmas have formed by differentiation and related processes, they rise to their own neutral buoyancy levels, forming channels above the main surge channel (Figs. 3.23, 3.27).