Annals of the Former World

by John McPhee

  Most volcanoes and related phenomena—most manifestations of the sort represented by the surface history of Yellowstone—are lined up along boundaries of the twenty-odd plates that collectively compose the earth’s outer shell. The plates, which are something like a sixtieth of the earth’s radius, slide around on a layer of the mantle hot enough to be lubricious. Where plates spread apart (the Red Sea, the mid-Atlantic), fresh magma wells up to fill the gap. Where plates slide by one another (San Francisco, Jericho), the ground is torn and walls collapse. Where plates collide (Denali, Aconcagua, Kanchenjunga), impressive mountains form. In collision, one plate usually slides beneath the other, plunging—in the so-called subduction zone—as much as four hundred miles. The material carried down there tends to melt, and to rise as magma, reaching the surface in volcanic form, as in the Cascade Range, the Andes, the Aleutians, and Japan. Yellowstone, with all its magmatic products and bubbling sulphurs, attracts special attention in the light of this story, because Yellowstone is eight hundred miles from the nearest plate boundary.

  When the theory of plate tectonics developed, it asked as well as answered questions—and not a few of the questions were inconvenient to the theory. Many had to do with volcanism. For example, why was the island of Hawaii pouring out lava in the dead center of the Pacific Plate? Similarly, if volcanoes were the products of subduction zones, where was the nearest subduction zone to the Tibesti Mountains of Saharan Chad? The Tibesti massif—a couple of thousand kilometres from the leading edge of the African Plate—consists of shield volcanoes like Mauna Kea and Mauna Loa. Where was the closest subduction zone to the chain of peaks that culminates in Mt. Cameroon, a stratovolcano fifteen hundred miles from the nearest plate boundary of any kind? Moreover, some of the fine old conundrums of geology—problems that antedated the plate-tectonics revolution—remained standing in its aftermath. What could explain the Canadian Shield? The South American Shield? The South African Shield? How could so much Precambrian rock lie close to sea level and not have been buried in a thousand million years? What, in recent time, had lifted the platform of the Rockies, causing their exhumation? Why were Love and I, there on the platform, not at sea level? What had lifted the Colorado Plateau, subjecting it to incision by canyon-cutting rivers? What explained flood basalts? Plate tectonics seemed to have no relevance to them. With plate theory, you would think you could predict the sedimentary history of continents, but you couldn’t. Why were continental basins—the Michigan Basin, the Illinois Basin, the Williston Basin—several kilometres deep? If you expect a shieldlike situation as the ultimate scene, what could explain these anomalous deep basins? Oil people wanted to know most of all. They asked plate theorists, “What does plate tectonics tell us about these basins?” The answer was “Nothing.” Why were the granites of New Hampshire relatively young, and therefore anachronistic in the Appalachian story? What explained great crustal swells, like random blisters on the ocean floor, rising high above the abyssal plains? What could explain Bermuda—a mountain summit seventeen thousand six hundred and fifty-nine feet above the Hatteras Abyssal Plain? What created the Marshall Islands, the Gilbert Islands, the Line Islands, the Tuamotu Archipelago—where corals veneer the peaks of twenty-thousand-foot mountains that tend to run in chains? Like Yellowstone, like Bermuda, like Hawaii, like Mt. Cameroon, they lie great distances from the nearest intersections of plates.

  Yellowstone draws its name from rich golden splashes of chemically altered volcanic rock. The place smokes and spits—the effects of proximate magma. On a geologic map of North America, Yellowstone appears at the eastern end of a bright streak of volcanic debris, coming off it like a contrail, extending across Idaho. With distance from Yellowstone, rock on that track is progressively older, descending age by age to the Columbia River flood basalts, which emerged from the ground like melted iron in early Miocene time, spread out across three hundred thousand square kilometres (in some places two and three miles deep), filled the Columbia Valley, and pooled against the North Cascades. By comparing the dates of the rock, one could be led to conclude that the geologic phenomenon now called Yellowstone has somehow been moving east at a rate of two and a half centimetres a year. As it happens, that is the rate at which, according to plate-tectonic theorists, North America is moving in exactly the opposite direction. In increasing numbers, geologists have come to believe that in a deep geophysical sense Yellowstone is not what is moving. They believe that the great heat that has expressed itself in so many ways on the topographic surface of the modern park derives from a source in the mantle far below the hull of North America. They believe that as North America slides over this fixed locus of thermal energy the rising heat is so intense that it penetrates the plate.

  The geologic term for such a place is “hot spot.” The earth seems to have about sixty of them—most older, and many less productive, than Yellowstone. Despite its position under thick continental crust, the Yellowstone hot spot has driven to the surface an amount of magma that is about equal to the over-all production of Hawaii, which has written a clear signature on the Pacific floor. Hawaii is the world’s most preserved and trackable hot spot. You can see its geologic history on an ordinary map if the map shows even the rudiments of what lies below the sea. The Pacific Plate is moving northwest. It dives into the Japan Trench, the Aleutian Trench, and regurgitates the volcanic islands that lie on the far side. The plate used to move in a direction closer to true north, but forty-three million years ago it shifted course. Any hot spot now active under the Pacific Plate will produce islands or other crustal effects that appear to be moving in the opposite direction—southeast. Mauna Kea and Mauna Loa—the shield volcanoes that from seafloor to summit are the highest mountains on earth—stand close to the southeasterly tip of the Hawaiian Islands. The extremely eruptive Kilauea is making the tip. The islands become lower, quieter, older—the farther they lie northwest. Islands older still—defeated by erosion—now stand below the waves. These engulfed ancestors of Hawaii form a clear track in the Pacific crust for more than five thousand miles. When their age reaches forty-three million years, their direction bends about sixty degrees to the north. Above the bend, they are known as the Emperor Seamounts. Ever older, they continue to the juncture of the Kuril and Aleutian Trenches, into which they disappear. The oldest of the Emperor Seamounts is Cretaceous in age. Mauna Loa, of course, is modern. Under the ocean forty miles southeast of Mauna Loa is Loihi, a mountain of new basalt, which has already risen about twelve thousand feet and should make it to the surface in Holocene time.

  The ages of the Emperor Seamounts and the familial Hawaiian islands create the illusion that Hawaii is propagating southeast at a rate of nine centimetres a year, while the message from plate tectonics is, of course, that the Pacific Plate is what is moving. The speeds and directions of the plates have been established by a number of corroborant observations. Offsets in faults like the San Andreas have been measured as expressions of time. Places in California that were once side by side and are now six hundred kilometres apart are also separated by eleven million years. A great deal of ocean-crustal rock has been dredged up and radioactively dated. The ages have been divided by distance from the spreading center to determine the rate at which the rock has moved. More recently, methods have been refined for making annual measurements of plate motions by satellite triangulation. Hot spots provide one more way of calculating plate velocities, for hot spots are to the drift of plates as stars to navigation.

  Conversely, it is possible to use established tectonic velocities to chart the tracks of hot spots with respect to the overriding plates. Given just one position and one date (the present will do), it is possible to say where, under the world, a hot spot would have been at any time across a dozen epochs. W. Jason Morgan, a geophysicist at Princeton, has sketched out many such tracks and reported them in various publications. Morgan can fairly be described as an office geologist who spends his working year indoors, and he is a figure of first importance in the history of the scien
ce. In 1968, at the age of thirty-two, he published one of the last of the primal papers that, taken together, constituted the plate-tectonics revolution. Morgan had been trained as a physicist, and his Ph.D. thesis was an application of celestial mechanics in a search for fluctuations in the gravitational constant. Only as a postdoctoral fellow was he drawn into geology, and assigned to deal with data on gravity anomalies in the Puerto Rico Trench. Fortuitously, he was assigned as well an office that he shared for two years with Fred Vine, the young English geologist who, with his Cambridge colleague Drummond Matthews, had discovered the bilateral symmetry of the spreading ocean floor. This insight was fundamental to the revolutionary theory then developing, and sharing that office with Fred Vine drew Morgan into the subject—as he puts it—“with a bang.” A paper written by H. W. Menard caused him to begin musing on his own about great faults and fracture zones, and how they might relate to theorems on the geometry of spheres. No one had any idea how the world’s great faults—like, say, the San Andreas and Queen Charlotte faults—might relate to one another in a system, let alone how the system might figure in a much larger story. Morgan looked up the work of field geologists to learn the orientations of great faults, and found remarkable consistencies across thousands of miles. He tested them—and ocean rises and trenches as well—against the laws of geometry for the motions of rigid segments of a sphere. At the 1967 meeting of the American Geophysical Union, he was scheduled to deliver a paper on the Puerto Rico Trench. When the day came, he got up and said he was not going to deal with that topic. Instead, reading the paper he called “Rises, Trenches, Great Faults, and Crustal Blocks,” he revealed to the geological profession the existence of plate tectonics. What he was saying was compressed in his title. He was saying that the plates are rigid—that they do not internally deform—and he was identifying rises, trenches, and great faults as the three kinds of plate boundaries. Subsequently, he worked out plate motions: the variations of direction and speed that have resulted in exceptional scenery. It was about a decade later when Morgan’s Princeton colleague Ken Deffeyes asked him what he could possibly do as an encore, and Morgan—who is shy and speaks softly in accents that faintly echo his youth in Savannah, Georgia—answered with a shrug and a smile, “I don’t know. Prove it wrong, I guess.”

  Instead, he developed an interest in hot spots and the thermal plumes that are thought to connect their obscure roots in the mantle with their surface manifestations—a theory that would harvest many of the questions raised or bypassed by plate tectonics, and similarly collect in one story numerous disparate phenomena.

  In 1937, an oceanographic vessel called Great Meteor, using a newly invented depth finder, discovered under the North Atlantic a massif that stood seventeen thousand feet above the neighboring abyssal plains. It was fifteen hundred miles west of Casablanca. No one in those days could begin to guess at the origins of such a thing. They could only describe it, and name it Great Meteor Seamount. Today, Jason Morgan, with other hot-spot theorists, is prepared not only to suggest its general origin but to indicate what part of the world has lain above it at any point in time across two hundred million years. Roughly that long ago, they place Great Meteor under the district of Keewaytin, in the Northwest Territories of Canada, about halfway between Port Radium and Repulse Bay. That the present Great Meteor Seamount was created by a hot spot seems evident from the size and configuration of its base, which is about eight hundred kilometres wide and closely matches the domal base of Hawaii and numerous other hot spots. If a submarine swell is of that size, there is not much else it can be. That it was once, theoretically, somewhere between Port Radium and Repulse Bay is a matter of tracing and dating small circles on the sphere traversed by moving plates.

  Keewaytin is in the center of the Canadian Shield. If the shield once had younger sediments on it, a hot spot underneath it would have lifted it up and cleaned it off, creating the enigma of the Canadian Shield. Morgan believes that various hot spots positioned in various eras under shield rock are what have kept it generally free of latter-day deposits. Stubborn fragments of the Paleozoic here and there on the shield suggest that this is so, as does the relatively modest number of meteorite craters. If the shield rock had been sitting there uncovered since Precambrian time, its surface could be expected to be more widely pockmarked, not unlike the plains of the moon.

  Later in the Jurassic, the Great Meteor Hot Spot was under the west side of Hudson Bay, and in the early Cretaceous under Moose Factory, Ontario. All this is postulated not on any direct field evidence but simply on a charted extrapolation from an ocean dome nearly four thousand miles away. As time comes forward, however, the calculations place the hot spot—with its huge volumes of magma—under New Hampshire a hundred and twenty million years ago. The radioactively derived age of a good deal of granite in the White Mountains, so puzzlingly “anachronistic” in Appalachian history, is a hundred and twenty million years.

  East of the North American continental shelf, lined up like bell jars on the Sohm Abyssal Plain, are the New England Seamounts. Their average height is eleven thousand feet. They are very well dated, and their ages decrease with distance east. Their positions and their ages—ninety-five million years, ninety million years, eighty-five million years—coincide with Morgan’s mathematical biography of the Great Meteor Hot Spot.

  A development that has greatly improved the precision of these measurements is argon-argon dating. A stream of neutrons in a nuclear reactor bombards a rock sample and causes a known fraction of its atoms of potassium to change into argon-39. Also in the sample are atoms of the isotope argon-40, which are unaffected by the bombardment and are the result of the natural decay of potassium through geologic time. The rate of decay is known and constant. The higher the proportion of argon-40, the older the rock. A mass spectrometer measures these ratios to establish a date. The older procedure known as potassium-argon dating—hitherto the best way of determining the age of something more than a few tens of thousands of years old—is done in two steps, requiring two samples. First, a chemical process determines how much potassium is present. Then a mass spectrometer looks at the second sample to see how much potassium has altered radioactively to become its daughter argon. The procedure suffers from the effects of weathering, which occur not only on the surface of rock but from grain to grain within. Argon slips away from weathered material, thus changing its over-all ratio to potassium and making any date determined by this method all the more approximate. Argon-argon dating is accomplished in the microscopic core of a single grain, beyond even the faintest disturbances of weather. The newer method is significantly more consistent and accurate than the older one. Results have shown—notably among the New England Seamounts—that where many potassium-argon dates fall into general approximation with Morgan’s calculations, the dates derived by argon-argon follow the track exactly.

  Eighty million years ago, in the Campanian age of late Cretaceous time, Great Meteor would have underlain the American-African plate boundary, the Mid-Atlantic Ridge. Since then, Great Meteor has cut a gentle curve southward through the African Plate. From late Cretaceous, Paleocene, and Eocene time, the path is as well defined as it is on the American side. After the Eocene, the hot spot made the big seamount that bears its name. Then it began to go cold, to evanesce, to fade like a shooting star.

  Shooting star. Almost everyone who describes hot spots is tempted to reverse reality and go for illusion at the expense of fact —that is, to narrate the apparent travels of hot spots as if they were in motion leaving trails like shooting stars, instead of telling the actual story of slow crustal drift over the fixed positions of thermal plumes. Myself included. With words, it is much easier to move a hot spot than it is to move a continent. Here, for example, is the story of another of the world’s hot spots told in terms of its illusory motion. With the flood basalts of Serra Geral, in southern Brazil, a hot spot is said to have begun in late Jurassic time. It moved east under Brazil for several million years and then cr
ossed over to Africa, which at that time was not much separated from South America. It lifted mountains in Angola, and then, doubling back, headed southwest under the ocean to form the Walvis Ridge, a line of seamounts leading to the hot spot’s present position—Tristan da Cunha.