Mantle Plumes and Mountains

J Brendan Murphy, Gary L Oppliger, George H Brimhall, Andrew Hynes. American Scientist. Volume 87, Issue 2. Mar/Apr 1999.

Traveling west on Sixth Avenue from downtown Denver, the Front Range of the Colorado Rockies looms 2,500 meters above the mile-high city. Yet as magnificent as the panorama may be, it fails to do justice to the Laramide Orogeny, the process that started the Rocky Mountains’ growth some 75 million years ago. Geologists estimate that the total uplift of the Front Range exceeded 7,000 meters. What drove the western half of Colorado to fracture and pile up to a height seven kilometers greater than that to the east? Earth scientists have long labored for a convincing explanation.

The theory of plate tectonics has offered clues. In a brief 30 years, it has revolutionized our understanding of mountain building. According to this theory the outermost layer of the earth, the lithosphere, is composed of a mosaic of rigid plates that ride on a hot, pliable layer of the earth’s mantle, the asthenosphere. As a consequence of circulation in the mantle, plates move with respect to each other at rates of a few centimeters per year. Over geologic time, this motion can account for the creation and destruction of oceans, the generation of mountain belts and sedimentary basins, the distribution of volcanic and earthquake activity, and the locations of ore, oil and gas deposits. Yet plate-tectonic theory has a tough time with the details of the Laramide Orogeny.

The conventional explanations of mountain building according to platetectonic theory all include horizontal plate motions and directly or indirectly depend on subduction zones, areas where oceanic crust descends back into the asthenosphere. Yet the Rocky Mountains lie 2,000 kilometers to the east of the current coastal margin, many times farther from an active subduction zone than mountain building generally takes place.

We propose that an additional mechanism of mountain building has been largely overlooked and may help explain not only the Laramide Orogeny but also other unusual geological features of the southwestern U.S. Our model involves the interplay of the horizontal motions of traditional subduction-related mountain building processes with vertical plumes of anomalously hot mantle ascending from thousands of kilometers below the earth’s surface. Together these mechanisms may offer a convincing explanation for what long has been a geologically puzzling part of the world and may lead to better understanding of mountain building worldwide.

Horizontal Forces In plate tectonics, as plates move apart, magma ascends from the asthenosphere, cools and solidifies to generate new lithosphere between the plates. This ongoing activity drives wedges of new lithosphere between the plates, separating them and generating a widening ocean. According to the theory, all oceans form in this way. The northern Atlantic Ocean, for example, has formed and progressively widened over the past 120 million years as the European and North American plates move apart. On a globe of constant radius, however, the divergence of plates and the construction of new lithosphere in some places must be compensated by convergence and destruction of lithosphere in others.

This is neatly accommodated by the recycling of oceanic crust in subduction zones. As it descends, the slab of lithosphere progressively heats up in the warmer ambient temperature of its surroundings. This eventually causes melting in the vicinity of the slab and in the overlying plate. These melts exploit weaknesses (such as fractures) in the overlying plate and ascend to the surface to produce volcanoes. In this way subduction leads directly to mountain building. The Andes are a modern example of mountains formed by such a process, and many of their highest peaks are either volcanically active now or have been in the recent past.

A second form of orogenic activity involves microcontinents, small islands or island chains located on oceanic crust. All modern oceans contain these islands; the Hawaiian chain is an example. Ultimately, when the tract of oceanic crust that separates these microcontinents from the continental margin becomes consumed by subduction, they will be swept to the margin and will collide with it. The impact results in deformation of rocks and igneous activity, which combine to form mountains. The coastal mountains of western North America formed in this way Repeated collisions with many of these small landmasses over the past 400 million years has caused the North American Plate to grow westward by an average of 500 kilometers, extending from Baja California to Alaska.

In some instances subduction and convergence consume an entire ocean, and two continental land masses collide, building mountains. Over the past 40 million years an ancient ocean called the Tethys was consumed by the collision of India with southern Asia and of northern Africa with southern Europe. The Himalayas and the Alps were pushed up in the Tethys’s place. None of these processes lends itself readily to an explanation of the Laramide Orogeny. Since the Laramide had no volcanic activity, conventional models of subduction do not apply. Likewise, it is clear that continents did not collide to form the Rockies. Furthermore, although collisions with microcontinents occurred during the time of the orogeny, these collisions were 1,200 kilometers away, at the very least, making this an unlikely explanation.

Geologists have been forced to invoke an unusual sort of subduction zone to explain the Laramide Orogeny one that had an extensive subhorizontal zone, rather than the more typical angled one. This zone must have extended at least 1,200 kilometers into the continental interior, and the oceanic crust must have been anomalously shallow in order to avoid melting and the generation of magma. Although this mechanism is widely accepted, the reasons why such a subduction zone should exist have been elusive. We may be able to fill in some of those details.

Plumes and Hot Spots

More than 30 years ago, Tuzo Wilson of the University of Toronto proposed mantle plumes to explain the formation of island chains. He suggested that Hawaii and several other volcanically active Pacific islands sit atop narrow columns, or plumes, of unusually hot rock and magma that ascend from deep within the earth. The interaction of plumes with the earth’s rigid outer crust causes broad bulges or swells, and the melting induced by plumes provides the raw material for some of the world’s most famous volcanoes. Volcanic centers above these plumes are known as hot spots.

In the example of the Hawaiian chain, the only active volcanoes are on Hawaii and the seamount Loihi, which is to the southeast of Hawaii and working its way toward the surface. Wilson noted that the islands are progressively older, less elevated and more eroded to the northwest along the length of the chain, and he interpreted this progression to be related to the westward motion of the Pacific Plate above “a jetstream of lava.” Each volcano was born in the present position of Hawaii directly above the plume. But as the plate moved northwestward, each was cut off from its supply of magma below. As each volcano cooled and aged, it subsided and became progressively more eroded.

Implicit in this analysis is the fact that hot spots are relatively stationary and certainly move more slowly than the plates above them. Building on this idea, Jason Morgan of Princeton University proposed that three parallel island chains in the Pacific Ocean could have been formed by the motion of the Pacific Plate over three different hot spots. Many investigators also think that the ascent of plumes is intimately associated with the breakup and dispersal of continents to form new oceans. Indeed, hotspot activity may have been an integral part of the breakup and dispersal of the supercontinent Pangea and the formation of the Atlantic Ocean. Don Anderson at the California Institute of Technology thinks that hot spots originate beneath large supercontinental land masses because continental crust conducts heat poorly compared with oceanic crust. By acting as an insulator, blocking the escape of heat from the mantle below, the supercontinent forces temperatures beneath it to rise, causing it to dome upward and eventually crack. Molten lava from the underlying asthenosphere rapidly ascends to fill the cracks, thereby driving the fragmented pieces of the former supercontinent farther and farther apart.

Nonetheless, hot spots are definitely not restricted to the locations of plate boundaries. The active Hawaiian volcanoes sit in the middle of the Pacific Plate at present, and mid-oceanic plate hot spots dot the globe. Thus their direct relationship to plate tectonics is unclear. Most earth scientists do accept that hot spots are the surface expression of hot columns of magma rising from a depth below the realm of plate tectonics, but just how deeply they originate is less certain. Recent evidence, however, suggests that these hot spots represent upwelling from near the core-mantle boundary, about 2,900 kilometers below the earth’s surface. (See American Scientist, March-April 1995, pp. 134147.) Thus plumes may be a phenomenon superimposed on plate motion rather than being a consequence of it. The tell-tale signs of the origin and ascent of such features were recently revealed by seismic tomography, a procedure analogous to computer-aided tomographic (CAT) scanning of the human body by criss-crossing waves from an x-ray generator. This technique, which combines information from a number of seismic waves emanating from earthquake zones that penetrate deep into the earth, allows the construction of a three-dimensional image of much of the inner earth.

Seismic tomography, combined with laboratory and theoretical models, provides insights into the geometry of plumes. An established plume has a relatively narrow central conduit in which hot mantle ascends, but it widens dramatically where it contacts the base of the lithosphere (see Figure 4). Evidence from Hawaii and from the Yellowstone Caldera, which also resides above a plume, indicates that the plume’s products are ponded at the base of the lithosphere and can “underplate” an area about 1,000 kilometers in diameter. This results in swelling and dynamic uplift of the lithosphere that is sustained as long as the plume remains active. Many of these regions are uplifted by as much as three kilometers above this hot, relatively buoyant material. Volcanic islands such as Hawaii rise more than another six kilometers above the dynamically domed sea floor, making them some of the highest mountains on earth.

Hot Spots and Mountain Building

As a plate moves over a hot spot, the crustal swell is dragged in the direction of plate motion into an eccentric elliptical shape up to nearly 2,000 kilometers in length. Hot spots and their associated crustal swells must inevitably interact with continental margins. More than 40 hot spots lie beneath the modern oceanic crust, and no modern ocean could be consumed without at least some hot spots being overridden by continental crust. In accordance with a basic principle of geology known as uniformitarianism, which states that modern processes are typical of those that have occurred throughout much of geological history, no ancient ocean could have been consumed without overriding ancient hot spots. Furthermore, the overriding of hot spots by continental margins must be a common phenomenon.

We think that the overriding of a hot spot and its large, buoyant, elongate swell by a convergent plate margin would dramatically affect the geometry of the subduction process and therefore would profoundly influence the style of mountain-building activity at continental margins.

A Modern Example

The geological evolution of the southwestern U.S. may offer such an example. From about 400 million to 75 million years ago this region experienced relatively normal mountain-building periods, including the Sonoma, Nevada and Sevier orogenies. Then, about 75 million years ago, a cycle of unusual tectonic processes began that continues to the present; we attribute these processes to the overriding of the Yellowstone hot spot by the North American Plate. Further, we conclude that the early manifestations of this event resulted in the Laramide episode of mountain building.

Before going into details, we must briefly review the geological history of western North America over the past 400 million years. During most of this time, mountain belts formed as the result of subduction and the episodic collision of microcontinents. Igneous rocks from that time have similar compositions to those in modern subductionzone regions, and the style and distribution of rock deformation is also typical of such settings. As oceanic crust was consumed by the subduction zone, microcontinents collided with the margin. Specific collisions produced important and discrete episodes of mountain building known as the Antler (about 400-330 million years ago), the Sonoma (260-210 million years ago) and the Sevier (120-70 million years ago) orogenies. In addition, from about 100 million to 75 million years ago, ongoing subduction resulted in the emplacement of large granitic bodies that make up the backbone of the Sierra Nevada.

The breakup of Pangea, which began 180 million to 150 million years ago, probably accelerated these processes. Since that time, the North American Plate has drifted westward as the Atlantic Ocean formed and has progressively widened. Thus the western margin of the North American continent now lies at a location formerly occupied by oceanic crust of a wider Pacific Ocean.

About 75 million years ago, an unusual succession of tectonic events began. Between about 75 million and 40 million years ago, there was widespread deformation as vast portions of the continental crust were tectonically sliced and heaved on top of one another in the Laramide Orogeny forming the early Rocky Mountains. The extent of this deformation (nearly 2,000 kilometers from the continental margin) is many times greater than the normal distribution of this type of deformation. During the same time, the region saw an almost complete lack of volcanic activity, a highly unusual situation for a typical subduction zone.

Both of these unusual features have been attributed to the presence of a subhorizontal subduction zone, rather than the more typical steeply angled zone that would have extended at least 1,200 kilometers into the continental interior. The deformation is associated with the interaction of this slab of oceanic crust with the overlying continental lithosphere. Because of the horizontal motion, the oceanic slab remained anomalously shallow and did not warm enough to generate magma. Although this scenario fits the available geological data, determining what would produce such a subduction geometry has proved difficult.

Equally enigmatic is the fact that after about 32 million years of quiescence, voluminous magma generation and associated volcanic activity began about 43 million years ago in northern Nevada and surrounding areas. This event may have had an important impact on modern economic development of the region. Many geologists think that this activity was directly responsible for gold mineralization in the area. Known as Carlin-type deposits (for their location near Carlin, Nevada), they qualify as one of the world’s most productive districts, having yielded 50 million ounces of gold ($18 billion at $365 per ounce) since their discovery in 1962. Another 150 million ounces may be accessible over the next 20 years.

Magmatic activity continued in the region until about 6 million years ago. Yet despite the fact that the western margin of North America was still converging with the Pacific Plate, geologists working the vicinity of this magmatism have found convincing evidence that the area was extending dramatically at the time. In the early 1970s, John Proffett, then a graduate student at the University of California at Berkeley, estimated that the crust in the area may have been extended by more than 100 percent, to at least twice its original width. This estimate has been supported by the more recent work of Brian Wernicke at Caltech and his colleagues. One manifestation of this extension is the dramatic block faulting of the Basin and Range Province. No consensus exists, however, on the underlying causes of the extension.

The chemical compositions of the igneous rocks generated at the initial stages of the magmatism 43 million years ago are typical of those produced by deep melting in the continental crust. They are relatively rich in elements such as silicon and form products such as granite and rhyolite. About 18 million years ago, however, major eruptions of basalt, a volcanic rock relatively poor in silicon and rich in iron and magnesium, began to accompany the more silicic eruptions. The chemical composition of these basalts bears a strong resemblance to that of the basalts of Hawaii, suggesting that they may be related to a plume beneath the region.

Over the past 16 million years, volcanism produced a pronounced linear trough known as the Snake River Plain, which stretches from the northern Nevada border in a northeasterly direction across Idaho toward Yellowstone National Park in northwestern Wyoming. The ages of volcanoes along the Snake River Plain become progressively younger moving from west to east, just as they do in the Hawaiian Islands. This age progression appears to be related to the southwesterly motion of the North American Plate above a stationary plume. Thus 16 million years ago northern Nevada was situated where Yellowstone is today, and it has moved progressively toward its present location since then.

The Model

Although evidence from the Snake River Plain indicates that the Yellowstone hot spot has existed for the past 18 million years or so, our model demands that it has existed at least for the past 75 million years. This is well within the typical life span of plumes such as the Hawaiian example, which shows no loss of strength after 76 million years. What might be the earlier history of the Yellowstone hot spot?

According to well-established plate reconstructions for the past 200 million years, the Atlantic Ocean lay much farther to the east 75 million years ago, and the western margin of the North American continent would have been 1,000 kilometers east of present-day Yellowstone National Park. Thus if the Yellowstone hot spot existed at that time and has indeed remained stationary since, it would have resided beneath an oceanic plate, as Hawaii does today If so, it would have generated similar ocean-island or seamount chains, the presence of which would be strong evidence for its antiquity.

The westward motion of the North American Plate and subduction along its continental margin would have destroyed direct evidence of the original Hawaiian-style island-chain geometry long ago. The subsequent motions of the islands, however, can be deduced. As previously discussed, such islands would have collided with the North American continental margin as microcontinents, a well known feature of orogenic activity along the western North American margin. Plate motions suggest that these collisions would not have been head-on; rather, they would have been highly oblique, resulting in deflection of the terranes northward along the margin. In fact, evidence of these former seamounts is found in the coastal regions of Washington and Oregon and as far north as the Yukon Territory of Canada.

Rocks in these regions bear all the hallmarks of their former oceanic Hawaiianstyle existence-including the nature of the volcanic eruptions, the chemical composition of the volcanic rocks and the presence of marine sediments. In addition, Stephen Johnson at the University of Victoria and colleagues working on the Yukon rocks were able to determine the latitude at which the volcanic rocks were formed. As iron-rich minerals crystallize out of magma and cool below 500 degrees Celsius, they behave as tiny magnets aligning themselves with the earth’s magnetic field. Because this direction depends on the latitude of the rocks at the time of cooling, the iron-rich minerals preserve information as they freeze that can betray the latitude at which they formed. The ancient latitude recorded in the Yukon rocks matches the latitude of the Yellowstone hot spot, providing strong evidence for its existence at least 70 million years ago.

Given its oceanic position 70 million years ago and its modern continental position beneath Yellowstone, the hot spot and its crustal swell must have been overridden by the continental margin in the interim. The actual position of the continental margin at each stage during the intervening period can be determined from published plate reconstructions, and a hot-spot track can be developed. This record, shown in Figure 7, indicates that the hot spot would have been beneath the continental margin 55 million years ago and under northern Nevada between 40 and 30 million years ago. It then tracked northeastward across Idaho to its present position beneath Yellowstone.

Although the commencement of the Laramide Orogeny 75 million years ago preceded the arrival of the hot spot beneath the continental margin, the crustal swell associated with the hot spot would have arrived sooner. Assuming a swell similar in size to that of the Hawaiian hot spot (elongated by 1,800 kilometers in the direction of plate motion), its arrival at the subduction zone may have preceded that of the hot spot by at least 15 million years, which is about the time the Laramide Orogeny began. The collision of this hot, buoyant, elevated oceanic lithosphere with the subduction zone provides an explanation for the onset of the Laramide Orogeny.

Recall that typical subduction zones are relatively steeply sloped because they subduct cold, dense oceanic crust, and the angle of descent into the asthenosphere is profoundly influenced by the density contrast between the crust and underlying mantle. In the vicinity of the hot-spot swell, the oceanic slab would have been considerably warmer and therefore more buoyant than normal, thereby producing an unusually shallow subduction. The buoyancy associated with progressive overriding of this swell may have aided in the gradual west-to-east shallowing of the subduction zone. This is consistent with geological data indicating that the age of deformation in the continental rocks is not uniform; rather it gets successively younger to the east.

The lack of magmatism can also be explained by the hot spot. Assuming that the hot spot was similar in dimensions to modern hot spots in the Pacific Ocean, we determined the approximate location 50 million years ago of an associated crustal swell similar in dimensions to that of Hawaii. The hypothetical location of the swell agrees remarkably well with the position of the “magmatic gap.” This suggests that the swell extended beneath the region of no magmatism and that its buoyancy contributed to the lack of magmatism.

As the North American Plate continued to migrate southwestward, the overridden plume would have been located beneath a progressively thicker cover consisting of the continental plate and the subducted oceanic slab. This additional cover would initially have reduced the ability of plume-generated melts to rise to the surface. In addition, the shallowness of the subduction zone would have further inhibited melting of the mantle. Therefore, magmatism associated with both the hot spot and with subduction would have ceased.

Calculations of the stability of modern oceanic plumes show that they have difficulty penetrating even thin oceanic crust, let alone thick continental crust. When the hot spot was first overridden by the North American Plate, it would have entered an “incubation” phase. Although hot mantle continued to be pumped under the lithosphere by the plume, this material could not ascend to the surface. Nevertheless, the modern record of volcanic activity near Yellowstone clearly demonstrates that the hot spot did eventually break through.

During the incubation period, the ponded plume progressively assimilated the overlying slab of subducted oceanic crust. About 40 million years ago, when the hot spot would have resided beneath the Battle Mountain region of northern Nevada, the plume made its first contact with the overlying continental crust. Since continental crust has a lower melting temperature than does oceanic crust, voluminous magmatism would be expected, as is indeed observed there. The rich Carlin-type gold deposits of northern Nevada may also be related to plume activity, since the plumes may have ascended from the core-mantle boundary, a region thought to be unusually rich in gold.

At the same time, the well-documented establishment of a more typical subduction zone at the continental margin is also consistent with the model. The assimilation and severing of the subhorizontal subducted slab by the hot spot would necessitate re-establishment of a new subduction zone at the continental margin to accommodate continued convergence between the North American and adjacent oceanic plate.

About 18 million years ago the style of magmatism changed from one dominated by silicic rocks generated by crustal melting to a geologically worldrenowned example of contemporaneous magmatism of two contrasting compositions. Although the silicic crustal melts persisted, they were accompanied by the eruptions that were poor in silicon and rich in iron and magnesium and that bear a strong chemical resemblance to those of another plume-the Hawaiian Islands. We argue that the voluminous basaltic volcanic rocks across the Columbia Plateau and the Snake River Plain were products of the Yellowstone Plume.

Finally, the rapid crustal expansion and magmatism in the Great Basin of Nevada that began about 30 million years ago has generally been attributed to the action of gravitational stresses associated with thick, weak continental crust. Our model, which places the Yellowstone hot spot beneath the Great Basin at the time of crustal expansion, provides a mechanism for generating this weakened crust.

The Fourth Mechanism

The presence of hot spots beneath oceanic lithosphere implies that they must play an important role in orogenic processes. The Late Mesozoic Cenozoic evolution of the southwestern U.S. may provide one example of the tectonothermal expression of such a process-with the Front Range being just one bit of evidence for it. Assuming that this model holds up to further scrutiny, we expect that many more examples will be found in the geologic record of the effects of horizontal plate motions combined with the vertical motions of mantle plumes. A few more pieces may thereby be added to the fascinating puzzle that is our planet.