THE BENIOFF ZONE is a sheet of earthquake foci that dips downward under an orogen at about 50 degrees to depths of about 300 km, starting from an oceanic trench. Andesitic volcanoes commonly occur above where the Benioff zone reaches a depth of 120 km. Further earthquakes may occur down to depths of as much as 700 km but these deep-focus earthquakes, although clearly associated, with the normal Benioff zone, appear to be somewhat independent of it: they only occur in the most active regions, so long stretches of Benioff zones do not have them. A discontinuity also occurs down the dip of the zone; even where deep-focus earthquakes are present a discontinuity o gap occurs below the normal Benioff earthquakes and the cluster of deep earthquakes. Incongruence may also be seen in plan view. For example, deep-focus earthquakes occur where expected down dip from the Bonin - Marianas trench, but the belt is less arcuate than the normal part of the Benioff zone, and continues on to cut across the orogenic arc and the trench at the southern end near Guam, and in the north continues roughly at right angles across the Honshu arc and trench and across the Sea of Japan, and after a short break reaches the Asian mainland near Vladivostok. This independence is strikingly apparent when the Earthquake distribution is studied stereoscopically from computer-generated plots. Many of the deep shocks show trans-current movement along the belt, which may account for its straightness. In contrast, the most active belt of deep-focus earthquakes anywhere conforms faithfully to the hooked pattern of the Tonga arc, although even there a pronounced gap is present between the normal Benioff zone of shallow and intermediate foci down to 300 km and the very numerous deep foci.
 

The expansion and subduction models agree that this is a zone of shearing fracture, with the orogen side moving relatively upward and over the oceanic side. This relative motion is confirmed by the sense of first motion deduced from the earthquakes that propagate from the fractures. But the subduction model claims that the oceanic lithosphere is thrusting downward beneath the orogen and has done so for thousands of kilometers. In contrast, in the expansion model, the oceanic crust is stationary and the Benioff zone is the boundary of the upward -thrusting diapir, which flares in a bell shape as it rises. The total upward motion in the center of the orogen is a hundred or so kilometers, not thousands, and the motion at the Benioff zone is only tens of kilometers at most. The question is which side moves the oceanic side or the orogenic side?

Continuous seismic profiling shows relatively thin regular sediments on the oceanic side. monotonously undisturbed for thousands of kilometers, right up to the trench. But as the trench is crossed, tectonic violence of all kinds erupts: seismicity, thrusting, and gross slumping toward the trench repeatedly triggered as the continuous over-steepening of the rising reaches slope instability. Heat outflow from the interior is consistently low on the oceanic side, but as the trench is crossed it more than doubles, and even increases locally up to tenfold. This situation is incompatible with the subduction model, in which the overall trench - orogenic arc system is where cold oceanic lithosphere is descending deep into the mantle. This requires that the total heat flux in the region should be much less than average. By contrast, the expansion model asserts that the whole orogenic zone is where hotter mantle material is forced upward and outward by the expanding interior; this requires that the total heat flux in the region should be significantly more than average. which is precisely what is found.

Every thing in the orogenic zone moves up. The isotherms are tens of kilometers higher than average. In the axial zone, volcanoes bring up lavas from the partially melted mantle; magmas rise to crystallise in plutons; crystalline metamorphic rocks originally deep below the floor of the geosyncline are forced up a dozen kilometers or more, to crop out high in the mountains; serpentinite and peridotites, originally even deeper below these crystallines, do likewise; gneiss domes intrude diapirically upward into the geosynclinal strata. These diapirs pierce upward in the viscosity pecking order displayed in Fig. 46-gneiss domes, migmatite diapirs, magmatic plutons, volcanic lavas, and nuees ardentes. Overthrusts are numerous, always with the orogen side riding up over the trench side, on very steep thrust surfaces near the axis of the orogen but on progressively flatter slopes farther from the axial zone. What forces all these bodies up? Not buoyancy. because except for the magmas they are denser than the rocks they penetrate. and they are driven to higher altitudes than their density would warrant. They are forced up by the ascending mantle below them which changes phase to less dense forms as higher temperatures are brought to levels of lower confining pressure.

From the axial zone right out to the Benioff zone is a single complex diapiric unit. The Benioff zone is the boundary of the ascending diapir against the stationary oceanic Lithosphere. The boundary of o a salt diapir is sharp because of the difference in the "Viscosity" of salt and intruded strata, but the boundary of an orogenic diapir is gradational, because the resistance to deformation ("viscosity") is not a discontinuous step but diminishes with temperature. The axial zone ascends fastest because the temperatures highest there, and this is maintained because the ascent itself continues to bring up hotter material. The pressure driving the diapir is great enough to fracture rock anywhere in the orogen, but in the central zone the rock yields by flow at lower stress-difference than the fracture stress; at lower temperature high in the orogen, thrust fractures even in the central zone. Laterally from the axial zone, stress-difference has to build up progressively higher to force flow as fast as the diapir requires, and eventually the stress-difference reaches fracture level relief by flow. The resulting zone of fracture is the Benioff zone. It is the ultimate boundary of the diapir. The distribution of earthquake foci reflects this pattern. At shallow depths earthquake fractures occur right across the orogen, but contract to the narrow Benioff zone progressively with depth.

Imagine a cylinder maintained at red heat pressed against a mild steel plate by a hydraulic ram. The steel in contact with the hot cylinder would slowly yield, but cracks would appear in the cold peripheral ring - the Benioff zone. Compare this with a bar of hot toffee, clamped horizontally in a vice, and with a heavy weight hanging on the projecting end. The toffee bar bends down immediately. Repeat: this with a similar bar, not quite so hot, and hang on the same weight. Again the bar bends, but more slowly. Repeat with a still cooler bar; again the bar bends down, but still more slowly. Repeat it with a cold bar and bending is scarcely noticeable before it snaps off, if the weight is greater than the strength. This is what happens at the Benioff zone, the outer boundary of the orogenic diapir.

Where the upward driving force is great. and the rate of ascent of the diapir relatively fast, the Benioff of fractures Goes deeper to higher temperatures. (A bar of warm toffee can be broken if bent rapidly, so that it breaks before it can bend.) Where the diapir driving force is less, the rate of upward movement can be accommodated by flow in relatively cooler rocks so the Benioff fractures only go down a couple of hundred kilometers, even though the ascending diapir originates much more deeply than this.

The rocks just outside (beneath) the Benioff zone are held in a state of elastic strain, at a stress level just below what would break them. They are the constraining wall of the diapir. Shear fracture could not occur without such a constraint, because an unconstrained shear stress produces rotation not shear fracture as in the Benioff zone. Experiments n rocks held under shear stress below fracture level show that they transmit sound and seismic waves 5 percent faster than when not stressed. The seismic velocity in the zone underneath the Benioff zone has been found to be higher than normal. This has been interpreted as evidence for a cold descending lithosphere slab, whereas such anomalous velocities are inevitable in the constraining rocks of the ascending diapir, just below the threshold of fracture.

In a salt dome, the outer boundary surface between the rising diapir and the strata that are only dragged upward, the surface on which the major stratigraphic discontinuity occurs, is the equivalent of the Benioff zone (Fig. 56).

Seismograms from a number of observatories of the same earthquake enable determination of the direction of motion at the point of rupture and differentiation between transcurrent, extensional, and shear failure. Except for shallow earthquakes, ruptures in the Benioff zone are shear failures. This has been interpreted as indicating crustal compression, incompatible with a generally extensional regime. How-ever, the bell-shaped boundary of a diapir must have this character, with the diapir side riding up over the static side. However, it has already been pointed out (Fig. 19 and associated text) that even in an extensional regime, tensional ruptures are only possible at shallow depths where the weight of the overburden is less than the shear strength of the rock, because at any greater depth shear rupture must occur before the regional extension can reduce any stress to zero to induce tensional failure.

Professor William Tanner, of Florida State University, has demonstrated both empirically and experimentally that diapirs only occur in extensional regions. The orogenic arcs of East Asia, particularly the Japanese arcs, which are the type areas on which the Benioff concept was founded, differ from the paradigm of Fig. 62 in three fundamental ways. Firstly, they are polycyclic, with a Tertiary orogenic cycle overprinted on a late Mesozoic cycle, along a zone which had been active far back in time to the Proterozoic birth of the Pacific. Secondly, the Mesozoic and Tertiary orogens were intrinsically asymmetric, in that the western flank was continental lithosphere, whereas the eastern side was oceanic. Thirdly, the rate of orogenic diapirism became exceptionally fast (Fig. 68).

The Orogenisis model of Fig. 62 starts with normal continental lithosphere which is progressively thinned to zero and replaced by the orogenic diapir. If the rate of ascent of mantle-derived material increases with further more rapid extension, the orogenic complex may be swept aside, leaving only this simatic material above the axial zone. The result is a nascent ocean basin with a central spreading ridge.

The bottom of the Benioff surface where it can be traced deeply enough lies under the small tensional seas (which 30 years ago I called "disjunctive seas" and recently were called by plate tectonicists "back- arc basins"), for example, the Sea of Japan. Professor Forese Wezel of Urbino has emphasised that the axis of the Cretaceous orogen lies here, but as the diapir continued to bring up material derived from the mantle, a stage was reached where the central zone of the orogen consisted only of mantle material, so that a new sea was born with a floor of mantle-derived (oceanic) crust. The earlier "continental" parts of the orogen were largely swept aside during the Palaeogene like surface scum over a convection cell, although residual horst-slices of it remain, as in the ridges in the center of the Sea of Japan. If the African rift valleys were expanded by such a process until oceanic floor widened between the outer rims, the Ruwenzori horst would be left as such a ridge in the center of the new sea.

Wezel has pointed out that Eocene to Oligocene tensional faulting dominates the coasts on both sides of these disjunctive seas, with Miocene molassic sediments unconformably overlying them. Because the Sea of Japan marks the swept central area of the diapir, the heat flow from the mantle is much higher there than normal, commonly in excess of 2.4 heat-flow units.

Recently a French oceanographic team led by Guy Pautot on the R.V. Jean Charcot has identified the extinct spreading axis for 500 km in the South China Sea by the inward-facing fault scarps (which trend N 50 degrees E +/- 10 degrees) and the outward tilting of the blocks.

As pointed out in connection with the Appalachian clastic fans (Fig. 66), an orogen does not grow uniformly along its length, but tends to form diapiric foci some 600 or 700 km apart. If conditions were symmetrical funnel-shaped diapirs would form ring-shaped orogens at the surface Wezel's krikogens). But the intrinsic asymmetry of the lithosphere has resulted in the tendency for the continental lithosphere to move westward with respect to the oceanic lithosphere. Hence, the spreading ridges did not grow symmetrically, because crustal increments were inserted on the west side of the spreading ridge, rather than equally on both sides, as in the standard plate tectonic model. The result is a line of basins, with rifted continental crust on the western side a basin floor of oceanic crust which has grown from west to east, and an orogenic arc on the eastern side.

Diapiric salt extrusion should ideally be symmetrical also, but the salt glacier of Fig. 53 only flows one way. Similarly, the spreading lobe of the Heide salt dome (Fig. 56) spreads only in one direction. We shall see later that there are more compelling fundamental causes of asymmetry related to the earth's rotation. With the exception of the Aleutians and Sunda arcs, which are convex equatorward for similar reasons, all Benioff arcs, including the Antilles and Scotia arcs, are convex eastward, and orogenic overthrusting tends to be eastward, not only in the East Asian orogens, but also in the Cordilleran orogen.

Intra-continental spreading ,such as the African rift valleys, continuing as the Red Sea, and the south Atlantic and south Pacific, are all essentially symmetrical because in these cases the controlling conditions were symmetrical: initial rift in cratonic continent, through to continent facing continent across oceanic crust with a median spreading zone.

From the foregoing it is clear that orogenic zones are genetically identical with misnamed "mid-oceanic spreading ridges " which are not always mid-oceanic (I will show later that they never were mid-oceanic in the North Pacific), and when joined with the orogenic belts are seen to be really circumcontinental.

The question is often asked, why do orogenic zones mainly occur at margins of continents? This concept itself is inherited from earlier times before the mobility of continents had been recognised. When Pangaea is restored, the great Tethyan orogen cuts the Pangaean megacontinent in half, and the Caledonian-Appalachian orogen developed within the contemporary continent to separate America and Africa-Europe. We will see later that the Cordilleran orogen was born as an intracontinental orogen to separate East Asia from North America, and Australasia-Antarctica from South America(see Fig. 97 in Chapter 22).

Once an orogenic zone has developed it remains a hotter weaker zone right through the mantle and subsequent expansion of the earth tends to be concentrated there. Thus the Mesozoic North Atlantic mainly followed the recently active Caledonian-Appalachian orogenic zone, although rotated from it by a transcurrent phenomenon which will be discussed later. The Cordilleran orogen was born about a thousand million years ago as a major rift zone separating Asia and North America and has continued to widen since, developing the Pacific Ocean. The Tethyan orogen has remained a weak zone, along which gross sinistral displacement has occurred (to be discussed later), but also along which the Mediterranean and Caribbean seas have opened.

Another question asked is why does the Atlantic have no current orogen if orogens and oceanic spreading ridges are genetically the same. The same question could be asked for the Southern Ocean between Australia and Antarctica. The answer lies in the rate of diapiric spreading.

We will see later that the rate of Earth expansion rapidly increased during the late Mesozoic and Tertiary. With such rapid expansion, the spreading zone is soon carried beyond reach of residual continental crust and rapid sedimentation, so that although the mantle diapirism proceeds exactly as before except that the rate is faster, the summit of the ridge rarely reaches sea level, so neither a nearby continent nor self-cannibalism contributes sediments. Vast outpourings of lava (all submarine), seismicity (greatly reduced because of the high temperature right to the surface), a typical seismic and gravity root (Fig. 64), deposition of base-metal sulfide ores, and topographic relief on a similar scale as on a continental orogen (Fig. 65) all occur. But there are no granites or andesites, which are typical of orogens adjacent to continents, because these require the presence of sediments and remnant continental rocks for their generation. The absence of a Benioff zone, or any earthquakes of intermediate focal depth, and shallow earthquakes mainly confined to transverse fractures, would seem to indicate that the expansion is largely filled by basaltic lavas.