Annals of the Former World

by John McPhee

  As Louis Agassiz discovered, if you set stakes in a straight line across a valley glacier and come back a year later, you will see the curving manner in which the stakes have moved. If you drive fence posts in a straight line across the San Andreas Fault and come back a year later, almost certainly you will see a straight line of fence posts—unless your fence is in the hundred miles north of the Cholame Valley. There the line will be offset slightly, no more than an inch or two. Another year, and it will have moved a little more; and a year after that a little more; and so forth. In its seven hundred and forty miles of interplate abrasion, the San Andreas Fault is locally idiosyncratic, but nowhere more so than here in the Central Creeping Zone. Trees move, streams are bent, sag ponds sag. In road asphalt, echelon fractures develop. Slivers drop as minigrabens. Scarplets rise. The fault is very straight through the Central Creeping Zone. It consists, however, of short (two to six miles), stepped, parallel traces, like the marks made on ice by a skater. Landslides occur frequently in the Central Creeping Zone, obscuring the fresh signatures of the creep.

  “The creep is relatively continuous for a hundred and seventy kilometres here and seems to account for nearly all of the movement,” Moores remarked. “Creep is rare. Most fault movement is punctuated. The creep produces numerous small earthquakes. There are actual ‘creep events,’ wherein as much as five hundred metres of the fault zone will experience propagating creep in one hour.” There were many oaks and few people living in the creep zone. The outcrops on the Pacific side of the fault sparkled with feldspar and mica—the granitic basement of the Gabilan Range. More than three thousand feet in elevation and close against the fault trace, the Gabilan Range creeps, too.

  Jumping and creeping, the San Andreas Fault’s average annual motion for a number of millions of years has been thirty-five millimetres. The figure lags significantly behind the motion of the Pacific Plate, whose travels, relative to North America, go a third again as fast. In the early days of plate tectonics, this incongruous difference was discovered after the annual motion of the Pacific Plate was elsewhere determined. The volcanic flows that crossed the San Andreas and were severed by the fault had not been carried apart at anything approaching the rate of Pacific motion. This became known as the San Andreas Discrepancy. If the Pacific Plate was moving so much faster than the great transform fault at its eastern edge, the rest of the motion had to be taken up somewhere. Movements along the many additional faults in the San Andreas family were not enough to account for it. Other motions in the boundary region were obviously making up the difference.

  With the development of hot-spot theory (wherein places like Hawaii are seen as stationary and deeply derived volcanic penetrations of the moving plates) and of other refinements of data on vectors in the lithosphere, the history of the Pacific Plate became clearer. About three and a half million years ago, in the Pliocene epoch, the direction in which it was moving changed about eleven degrees to the east. Why this happened is the subject of much debate and many papers, but if you look at the Hawaiian Hot Spot stitching the story into the plate, you see, at least, that it did happen: there is an eleven-degree bend at Pliocene Oahu.

  The Pacific Plate, among present plates the world’s largest, underlies about two-thirds of the Pacific Ocean. North-south, it is about nine thousand miles long, and, east-west, it is about eight thousand miles wide. What could cause it to turn? Various events that occurred roughly three and a half million years ago along the Pacific Plate margins have been nominated as the cause. For example, the Ontong-Java Plateau, an immense basaltic mass in the southwest Pacific Ocean, collided with the Solomon Islands, reversing a subduction zone (it is claimed) and jamming a huge slab of the Pacific Plate under the North Fiji Plateau. The slab broke off. Suddenly released from the terrific drag on its southwest corner, the rest of the great northbound plate turned eleven degrees to the northeast. A number of coincidental collisions along the plate’s western margin may have contributed to the change in vector. Additional impetus may have been provided by the subduction of a defunct spreading center at the north end of the plate. The extra weight of the spreading center, descending, may have tugged at the plate and given it clockwise torque. Whatever the cause, it’s not easy to imagine a vehicle that weighs three hundred and forty-five quadrillion tons suddenly swerving to the right, but evidently that is what it did.

  The tectonic effect on North America was something like the deformation that results when two automobiles sideswipe. Between the Pacific and North American plates, the basic motion along the San Andreas Fault remained strike-slip and parallel. But as the Pacific Plate sort of jammed its shoulder against most of California a component of compression was added. This resulted in thrust faults and accompanying folds—anticlines and synclines. (Petroleum migrated into the anticlines, rose into their domes, and was trapped.)

  Earlier—about five million years before the present—the ocean spreading center known as the East Pacific Rise had propagated into North America at the Tropic of Cancer, splitting Baja off the rest of the continent, and initiating the opening of the Gulf of California. The splitting off of Baja was accompanied by very strong northward compression, which raised, among other things, the Transverse Ranges above Los Angeles, at the great bend of the San Andreas. That the Transverse Ranges were rising compressionally had been obvious to geologists long before plate tectonics identified the source of the compression. But not until the late nineteen-eighties did they come to see that compression as well as strike-slip motion accompanies the great fault throughout its length, as a result of the slight shift in the direction of the Pacific Plate three and a half million years ago. All these compressional aspects taken together—anticlines, synclines, and thrust faults in a wide swath from one end of California almost to the other—account for some of the missing motion in the San Andreas Discrepancy. The Los Angeles Basin alone has been squeezed about a centimetre a year for two million two hundred thousand years. The sites of Laguna Beach and Pasadena are fourteen miles closer together than they were 2.2 million years ago. This has happened an earthquake at a time. For example, both the Whittier Narrows earthquake of 1987 and the Northridge earthquake of 1994 lessened the breadth of the Santa Monica Mountains and raised the ridgeline.

  The Whittier Narrows hypocenter was in a deeply buried fault in a young anticline. Such faults tend to develop about ten miles down and gradually move toward the surface. Northward for five hundred miles, young anticlines on the east side of the San Andreas Fault are similar in nature—the products of deep successive earthquakes. Most are recently discovered, and many more, presumably, remain unknown. They make very acute angles with the fault, like the wake of a narrow boat. When a temblor goes off like a hidden grenade, geologists often have not suspected the existence of the fault that has moved. The 6.5 earthquake at Coalinga in 1983 was that kind of surprise. It increased the elevation of the ridge above it by more than two feet.

  In 1892, a pair of enigmatic earthquakes shook Winters, which is near Davis, in the Great Central Valley. Evidently, the earthquakes occurred on the same sort of blind thrust that is under Coalinga, but the Winters thrust is of particular interest, because it is east of the Coast Ranges and fifty miles from the San Andreas. Yet it is apparently a product of the newly discovered folding and faulting that everywhere shadow the great fault. The Central Valley of California is about the last place in the world where virtually any geologist would look for an Appalachian-style fold-and-thrust belt. Without shame, Moores sketches one on a map of California; it goes up the west side of the valley almost all the way from the Tehachapi Range to Red Bluff and reaches eastward as far as Davis. He and his Davis colleague Jeff Unruh have been out looking for tectonic folds in the surreally flat country surrounding the university. This is a game of buff even beyond the heightened senses of the blind. They have found an anticline—an arch with limbs spread wide for many miles and a summit twenty-five feet high. They call it the Davis Anticline. It is a part of what Moores likes to desc
ribe as “the Davis campus fold-and-thrust belt.” He is having fun, but the folds are not fictions. The anticline at Davis has developed in the past hundred thousand years. It is rising ten times as fast as the Alps.

  On perhaps the weirdest geologic field trip I have ever been invited to observe, he and Unruh went out one day looking for nascent mountains in the calm-water flatness of the valley. There were extremely subtle differentiations. Moores said, “We are looking here on the surface for something that is happening five kilometres down—blind thrusts. Compressional stress extends to the center of the valley.”

  “Topography doesn’t happen for nothing,” Unruh said. “Soil scientists have long recognized that these valley rises are tectonic uplifts. Soils are darker in basinal areas. There’s a fault-propagation fold in this part of the valley.”

  Moores later wrote to me:

  We continue to gather evidence. We have seen two seismic profiles that show a horizontal reflection, presumably a fault, that extends all the way from the Coast Ranges to the Sacramento River. Jeff has been working at stream gradients. The rationale is that where there is a sharp change in gradient on a flood plain there is a reason, and the reason here is uplift. The analysis fits the two areas of acknowledged uplift west of Davis pretty well, and seems to indicate a new north-trending zone of uplift that goes right through Davis itself. Maybe there was a reason why the Patwin Indians selected this particular spot on the banks of Putah Creek for this village, after all. It was a high spot in a swamp, and it was high because it is coming up!

  The compressive tectonism associated with the plate boundary contributes to the total relative plate motion, but not much: the over-all average is less than a centimetre a year. And that does not nearly close the numerical gap. Surprisingly, the rest of the missing motion seems to come from the Basin and Range, the country between Reno and Salt Lake City, wherein the earth’s crust has been stretching out and breaking into blocks, which float on the mantle as mountains—where the stretching has increased the width of the region by sixty miles in a few million years. Very-long-baseline interferometry has shown that the Basin and Range is spreading about ten millimetres a year in a direction west-northwest. This supplies enough of the total plate-boundary motion between the Gulf of California and Cape Mendocino to make up the difference in the San Andreas Discrepancy. If some Pacific Plate motion is coming from Utah, Utah is a part of the plate boundary.

  The westernmost range of the Basin and Range Province is the Sierra Nevada, which has risen on a normal fault that runs along the eastern base of the mountains. The fault has experienced enough earthquakes to give the mountains their exceptional altitude. The most recent great earthquake there was in 1872. In a few seconds, the mountain range went up three feet. In the same few seconds, the Sierra Nevada also moved north-northwest twenty feet. That would help to fill in anybody’s discrepancy.

  Perhaps a sixth of the total motion between the plates is contributed by the other faults in the San Andreas family. Each is strike-slip, active, right-lateral—that is, viewed from one side of the fault, the other side appears to have gone to the right.

  In a general way, you can demonstrate their relationship to one another with a deck of cards. Hold the deck, side up, between the palms of your hands, and slide the hands, pulling the right side toward you, pushing the left side away, and keeping pressure on the deck. The cards will respond by slipping, sticking, locking, sliding. Some may slide more than others. There may even develop a primary break. In any case, the fifty-one slips between the cards are, as in California, a family of right-lateral strike-slip faults. If one has moved more than the others, in effect you may have cut the cards, and you could call that cut the San Andreas Fault. But all the cards, to varying extents, have contributed slip to the total motion.

  Moores believes that the plate-vector change three and a half million years ago is what probably created so large a grouping of boundary faults. Parallel and subparallel to the San Andreas, they have been likened to the tributaries of a river or the branches of a tree. But they are not dendritic. Often they do not conjoin. They are more like the checks that appear in dry timber. They all trend northwest. Many of them have varying local names, because the field geologists who did the naming, as much as a hundred years ago, did not suspect their continuity. The actively slipping Green Valley Fault, which intersects I-80 at Cordelia, in the eastern Coast Ranges, continues to the south as the Concord Fault. I-680, branching off, follows the fault and stays right on it. Down through the Napa Valley and under San Pablo Bay and through Berkeley and Oakland and Hayward and farther south than San Jose runs a continuous fault that is in segments named Healdsburg, Rodgers Creek, and Hayward. Portentous microquakes on the Rodgers Creek Fault have suggested to the United States Geological Survey the possibility of an earthquake equal in intensity to the 1989 wrench on the San Andreas near Loma Prieta.

  The Calaveras Fault runs close to the Hayward Fault and extends somewhat farther south, like the Sargent Fault, the Wildcat Fault, the Busch Ranch Fault. The Antioch Fault is in the Great Central Valley. In seven hundred miles of splintery faults, the ones I have mentioned are all in the San Francisco Bay Area east of the San Andreas. West of it—on the San Francisco Peninsula or under the ocean—are the Pilarcitos Fault, the La Honda Fault, the Hosgri Fault, the San Gregorio Fault. The San Gregorio Fault extends from San Francisco to Big Sur, south of Monterey. Its longest historical jump is thirteen feet. It has produced great or major earthquakes on an average of once every three hundred years—nothing to be concerned about, unless it is your year. In the San Gregorio fault zone, San Mateo County has forty-acre zoning.

  A comparable cross-sectional anatomy of the San Andreas system could be described for any latitude from Cape Mendocino to the Salton Sea, with a long list of names of contributive strike-slip faults. Within the system, the twentieth-century earthquake second in severity occurred in 1952 on the White Wolf Fault, near Bakersfield. In southern California, the belt is as much as a hundred and fifty miles across—three times as wide as it is at San Francisco, and a good deal more complex.

  The Hayward Fault alone has contributed more than a hundred miles of offset. Running southeast from San Pablo Bay, north of San Francisco, to the latitude of Santa Cruz, it disappears near Gilroy, not far from the San Andreas Fault at San Juan Bautista. In many places, such as Berkeley, the Hayward Fault has Jurassic rock of the Franciscan melange on one side of it and Cretaceous rock of the Great Valley Sequence sort of dredged up on the other side. A return specialist in the football stadium going eighty yards through a broken field will gain or lose about fifty million years. In a split, unpredictable second, he can be tossed out of bounds by a shift of the sod beneath him. The Hayward Fault runs not only through Memorial Stadium but also through or very near the Alameda County hospital, the San Leandro hospital, and California State University, Hayward. The Hayward Fault also ran through the California School for the Deaf and Blind, but the State became nervous, moved the school to another site, and then filled up its old dorms with Berkeley undergraduates.

  The Hayward Fault separates the Cretaceous Berkeley Hills from the Jurassic university campus and the Holocene alluvium of the flat ground near San Francisco Bay. To no small extent, the Hayward Fault has created the Berkeley Hills, which are an obvious fault scarp, the result of a vertical component in an otherwise strike-slip motion. The change is abrupt from gentle slope to steep escarpment because the fault is so active and the hills are so young. Large earthquakes occurred on the Hayward Fault in 1836 and 1868—not the sort of information that is likely to plumb the tilt in a laid-back sophomore at Berkeley. However, a U.S.G.S. Miscellaneous Field Studies map of predicted maximum ground-shaking from large earthquakes on the San Andreas Fault and the Hayward Fault shows three areas of A-level intensity—characterized as “very violent”—for a Hayward earthquake: the environs of lower University Avenue in Berkeley; the blocks just east of Lake Merritt in downtown Oakland; and the Warren Freeway in Pi
edmont. The Warren Freeway uses the Hayward Fault in the way that water uses a riverbed.

  The Geological Survey sees a sixty-seven-per-cent chance of another major San Francisco Bay Area earthquake on either the San Andreas or the Hayward Fault before the year 2020, with probabilities leaning toward the Hayward, because a jump there is so long overdue. If it should equal the intensity of several nineteenth-century shocks, the new one—according to estimates by the Federal Emergency Management Agency and the California Division of Mines and Geology—would result in as many as four thousand five hundred deaths, a hundred and thirty-five thousand injuries, and forty billion dollars’ worth of damage. The Geological Survey adds a comment to these possibilities: the center of San Francisco “is as close to the Hayward Fault as it is to the San Andreas Fault.”

  San Lorenzo Creek, coming out of the hills into the outskirts of Hayward, bends sharply right when it hits the fault, flows northwest on the fault for more than a mile, and then takes a left and heads southwest to the bay. Hayward is about fifteen miles down the trace from Berkeley, and if ever there was a type locality to define the geological meaning of “type locality” Hayward is the place. Rarely is a fault zone so sharply drawn. Through part of the town runs a steep emphatic bench, resembling a sloped medieval wall, that is the product of the fault beside it. On D Street between Mission and Main, the curb, running east, makes a right-lateral bend to the south, and then continues east. When Moores and I were on D Street not long ago, a pressure ridge had buckled the sidewalk. The fault went through the service department of Boulevard Buick. There were fresh patches in the sidewalk outside. Between C and D, a building that had long housed the municipal government—and had been designed and constructed as the Hayward City Hall—stood precisely on the fault, which was tearing the building apart as if it were a tuft of cotton. Tiles and plaster had showered the bureaucrats until they fled. The bureaucracy included a department established to deal with emergencies confronting the city. At 934 C Street was a store that declared itself to be the “Hayward Sewing Center, Alterations.” The alterations included a canted curb, a pressure ridge in the sidewalk, and a wall in the act of bulging toward Mexico City. There was a bend in the long wall of the Action Signs Building, at 22534 Mission. The fault had offset the Spoiled Brat Parking Lot. In a municipal parking lot nearby, the lines of meters took right-lateral bends and then resumed their original direction. Robert’s School of Karate—in motion relative to the antique shop next door—had recently moved south an inch. A sign in the window said “KARATE KENPO GUNG-FU BOXING—The Only School in Hayward Teaching Street Fighting.” Sidewalks were patched on every east-west street. The occupied house at 923 Hotel Avenue had been so torqued by the fault that one of its walls was concave. Looming over it—fifty feet high—was the Hayward escarpment.