The ABC's of Plate Tectonics

The Mechanism of Plate Tectonics

by Donald L. Blanchard

What does appear to be accepted is that plate movement is driven by convection within the upper mantle, powered by the temperature difference between the (relatively) cold surface and the heat generated deep within the Earth. Whether the plates are themselves the upper part of this convection system, with a static or reverse-flowing fluid mantle beneath, or whether they are just rafted along on the surface of an actively convective mantle, remains unanswered.

Clearly, for the plates to move, they must either be pushed away from spreading centers, pulled down into trenches, or dragged along by the friction of convection on their undersides. Conventional wisdom has it that all three contribute, although the proportions contributed by each force remains undecided.

Ridge Push is the force applied to plates at spreading centers. As plates separate, new, hotter material is extruded into the gap between plates. The elevated temperatures lower the density of the recently deposited material, causing it to float higher in the mantle; hence the formation of ridges. Gravity then takes over, drawing the ridge material down and away from the spreading center, thereby widening the gap for more hot mantle material to upwell.

Slab Pull is considered by many to be the dominant, or even the only, driving mechanism for plate tectonics. Subduction zones typically occur a long distance away from spreading centers, where the plates have had plenty of time to cool off. Lower temperature makes the plate more dense than the material beneath it, causing it to subduct into the mantle. As the plate edge is drawn under, it pulls the remainder of the plate behind it.

Plate Drag is supported in several different models of plate tectonics. In one model, the upper mantle is separated from the lower mantle, with little or no transfer of material between them. Small, local convection cells in the upper mantle can occur beneath the interiors of plates and can contribute frictional drag forces to reinforce plate movement. In another model, the upper and lower mantle are more tightly coupled, with material from plumes originating deep within the lower mantle and contributing significantly to upper mantle convection, and hence to plate movement.

Ridge push and slap pull can explain why plates remain in motion once spreading is initiated, but do not adequately explain what initiates a rifting episode in the first place. The traditional wisdom is that new rifts form under large continental masses, such as supercontinents, where the greater insulating properties of continental material cause an increase in heat buildup, leading to uplift, softening, and ultimately rifting of the continental material. It is true that the supercontinent of Pangaea was incredibly short lived, on a geologic time scale, starting to break up almost as it was being formed. (See the companion article "The Formation of Pangaea: The Making of a Supercontinent", also on this site.) But Gondwanaland, which was a supercontinent in its own right, survived for almost a billion years before it broke up during Pangaea's formation. And there was no supercontinent, in fact no continent of any sort, over the East Pacific Rise 40 million years ago when it began spreading in a northwest-southeast direction.

Clearly the force applied to move the plates must be capable, at least occasionally, of exceeding the tensile strength of the plate material and of breaking a larger plate into two smaller ones. If this were not the case, the Earth would be covered by a single spherical plate, and no relative motion between plates could occur. Also, the force applied cannot normally be greater than the tensile strength of the plates, or else small pieces would be breaking off all the time, and there would be a myriad of very small plates bobbing about, rather than the very few, large plates that we know to exist.

The obvious cantidate for a mechanism for transforming thermal energy into the movement of plates is the action of plumes. Plumes are thought to be columns of heated mantle material rising from deep within the mantle, possibly from the very bottom of the mantle where it contacts the Earth's liquid core. The problem with models using plumes to power plate movements is that the volume of material rising in the plumes is thought to be small compared to the volume of the moving plates themselves. Yet if the plume material is hot enough, which it shows every evidence of being, there may well be enough energy available.

The Earth's mantle is not composed of a single mineral, with a sharp melting/freezing point. Rather, it is a mixture of a large number of different substances, each with a different melting temperature. Thus while the upper surface of an oceanic plate is a crystalline solid, the underside exhibits a gradual phase change, from solid to plastic to viscous liquid as the temperatures and pressures increase with depth. Yet even at great depth, the mantle remains an exceedingly viscous fluid, flowing ever so slowly even in response to extreme pressures. Any force postulated to either push or pull oceanic plates would be quickly dissipated overcoming the friction of sliding over a static fluid mantle beneath.

Furthermore, any pulling force great enough to tear a new plate off from an existing one would surely tear off a piece immediately adjacent to where the pulling force is applied (and therefore is the greatest, undiminished by friction), rather than creating a new rift somewhere clear across the ocean. Nor can slab pull explain the orogenies resulting from continental collisions. Once two continents meet, any oceanic slab separating them will have been totally consumed by subduction, and the plate's leading edge will be made up exclusively of continental material, which is too buoyant to be pulled down into the mantle. Yet the Indian plate has continued to drive northward into Asia for something on the order of a thousand miles (1600 km) since its collision with Asia. Only push or drag forces can explain the Himayalas.

Similarly, any pushing force great enough to buckle a plate and form a new trench and subduction zone would almost certainly form one immediately adjacent to an existing spreading center, or at least no farther away than the point at which the new plate has cooled to a normal deep ocean temperature, rather than at the opposite side of the ocean or the middle of a continent. These arguments support the notion of a convective mantle, with the plates being rifted apart and transported by the accumulated forces of frictional drag on their undersides.

It should be noted that it is not necessary to postulate that the convective current beneath a mobile plate exactly match the movement of the plate above. Any convection cell model for the mantle that involves horizontal flow in many different directions beneath the plates will cause motion of an overlying plate; the plate will simply move as the vector sum of all the drag forces applied to its underside, provided that the frictional drag forces applied do not exceed the shear strength of the plate material.

Add multiple heat sources of equal intensity, uniformly spaced beneath the fluid, and multiple toroidal convection cells, all of equal size, will form. If the heat sources are of unequal size, hotter sources will create larger toruses. This provides the basis for the proposed model for convection in the Earth's mantle, with the heat sources lying deep within the Earth and plumes being the rising columns at the center of (approximately) toroidal convection cells.

Plumes are scattered in an apparently random manner all over the surface of the globe. Many, but not not all, of the larger plumes are situated along spreading centers, suggesting that plumes may be instrumental in the formation of rifts and the origin of new spreading centers. A list of some of the more prominent currently active plumes is provided in the following table:

The material in the center of the column, being the hottest, will rise fastest, and only this hottest material will likely reach the surface. When it does, it will spread out radially beneath the frozen plate material, while the cooler entrained material will flow outward beneath it. Surface flow under the plates will thus form a thermal gradient, with the hottest, fastest flowing material nearest the surface and progressively cooler and slower moving material flowing beneath it. Only when the surface material cools below the material beneath it will it descend into the mantle below, ultimately to return to the heat source.

Plumes rising from less intense hot spots may well encounter hotter material from a more intense adjacent plume spreading out above it, well before the lesser plume reaches the surface. Its material may then never reach the surface at all, but instead be deflected laterally to flow along with the material from the hotter plume, thereby enhancing the already existing flow. This is, of course, pure speculation, and the degree to which such lesser plumes contribute to the movement of plates cannot be reasonably appraised. As their flow never reaches the surface, their contributions cannot be easily detected.

A possible convective model for how two plumes can act together to create a rift can be illustrated using as an example the Iceland and Azores plumes, which lie roughly on a north-south line. As the hot material from the Iceland plume reaches the surface, it spreads out radially in all directions, gradually cooling as it goes. However, the material that flows south soon runs in to the material flowing north from the Azores plume, well before either flow has cooled sufficiently to sink back into the cooler mantle material below. Where the two flows meet, both must therefore be deflected to the sides. As a result, along a line between the two plumes, all surface flow should therefore be redirected to the east and west, perpendicular to and away from the connecting line. This 'shear line' of opposing flows should effectively concentrate the drag exerted on the underside of the plates to a degree that may be sufficient to initiate, or at least, redirect a rifting episode. (The rifting episode that produced the North Atlantic appears to actually have started far to the west, first separating what later became southern Mexico from South America, and only later turning northward to open the North Atlantic.) A similar shear line, also essentially north-south but offset far to the east, seems to have been created when the Ascension and Tristan de Cunha Island plumes conspired to separate South America from Africa.

The Atlantic ocean continues to widen, and North America is being dragged westward. However, western North America seems to be moving westward faster than the rest of the continent. A succession of paired mountain ranges, their steeper sides facing each other and rift valleys separating them, progress eastward from California into Wyoming, Colorado, and New Mexico. Starting with the Sierra Nevada and White Mountains, separated by the Owens Valley, and the Panamint and Amargosa Mountains, separated by Death Valley, paired ranges progress across Nevada and on into the Rocky Mountains of Wyoming and Colorado. The easternmost pair of ranges in Wyoming are the Tetons and the Absaroka/Wind River Range, separated by Jackson's Hole, a rift valley whose northern end is Yellowstone National Park, site of the Yellowstone plume. In southern Wyoming and Colorado, paired ranges, offset considerably to the east, include the Medicine Bow and Laramie Ranges, separated by North Park; the Never Summer, Gore, and Front Ranges, separated by the upper Colorado River Valley, Middle Park, and the Arkansas River Valley; and the San Juan and Sangre de Christo Mountains, separated by the San Luis Valley. The rift zone ends in the Rio Grande rift valley, near a minor plume identified by heat flow measurements near Socorro, New Mexico.

The model suggested for these features is that a shear line exists between the Yellowstone and Socorro plumes. However, the surface flow of plume material from this line is insufficient to overcome the westward flow from the Mid-Atlantic shear line, driven by the Iceland and Azores plumes. Rather, all of the flow from the Yellowstone/Socorro shear line is deflected westward by the more powerful currents from the east. Thus all of North America is being dragged westward, but the western part is being pushed harder than the rest. As a result, a rift forms wherever the lesser shear line occurs, but continued westward drift soon pushes the incipient rift west of the shear line, where it becomes inactive, and a new rift is initiated farther to the east. The current series of rifts between Yellowstone and Socorro are probably doomed to failure as surely as were their predecessors farther west.

The classic triple junction, however, is the Afar Triangle of East Africa. The Afar Triangle is actually a section of sea floor uplifted by the high heat flow from the plume beneath it. It sits at the junction of three currently active rift zones: the Gulf of Aden, the Red Sea, and the East African Rift Valley. At the opposite end of the Carlsberg Ridge - the spreading center responsible for opening the Gulf of Aden, is a plume beneath the Chagos Archipelago. This is the same plume that created the Maldive and Lacadive Islands farther north, and deposited the flood basalts of the Deccan Traps of India, when that subcontinent passed over the plume around 65 million years ago.

Two other plumes are also postulated by the proposed model; one under East Africa near the south end of the East African Rift Valley, which, it is suggested, is responsible for the volcanos Mt. Kilimanjaro and Mt. Kenya, and the other, perhaps a minor plume, near the north end of the Red Sea or in the Eastern Mediterranean Sea. The proposal is that a triple junction occurs where four plumes occur in close proximity to one another, with the three outer plumes forming a 'Y' and the fourth plume in the center. If two plumes are insufficient to initiate a rifting episode (and it may be that the 'shear line' proposed above is only sufficient to redirect and perpetuate a rift, but not to initiate one), then surely four plumes in close proximity should suffice.

Triple junctions are rather rare occurrences in the geological record, but the small number of plates into which the Earth's crust is divided testifies that the separation of plates and the formation of new spreading centers is likewise rare. We are fortunate indeed that such an event is playing itself out even as we watch.

Those who are interested are invited to also read the companion article:
"The Formation of Pangaea: The Making of a Supercontinent."

The opinions expressed in these lessons are those of their author, Donald L. Blanchard, and probably (in fact quite likely) do not coincide with those of any professional authority on plate tectonics or paleogeography.

| Return to First Page |
| Earth Sciences Home Page |