“Where, dammit, where?”
News from everywhere else poured from the radio, but there was no further mention of the quake. Karen was still fumbling with the tuning knobs as she pulled up in front of the Caltech lab and abandoned the car to race inside to the seismograph drums.
All the needles had jumped toward the mechanical stops with the wave arrivals of a genuine great quake, and initial calculations (which were already under way when she burst through the door) seemed to indicate that the tremendous pulse of seismic energy was coming from south of Mexico City, from the southern coast of Mexico.
Karen’s first suspicion was exactly right. It was indeed her earthquake!
A team of Texas seismologists had first indentified the area as ripe for a seismic “event” a year before in 1977, predicting a 7.25 to a 7.75 great quake within one and a half years. But she had been the first to build a net of seismographs in an attempt to “trap” the quake they thought was coming. Now here it was, roaring in on schedule (nine months into an eighteen-month forecast) and at an Ms magnitude of 7.8, which was on the upper end of their magnitude prediction.
Karen McNally, Ph.D., senior research fellow at Caltech and one of the brightest young seismologists on the trail of earthquake prediction, was ecstatic—and eager to get right back to Los Angeles International Airport and aboard any flight that could carry her south once again. Her paper could wait.
Something had been missing from the seismic puzzle of Mexico’s southern coast. The fact that a certain area near Oaxaca had not suffered a major earthquake since 1928 had been noticed by seismologists at the Geophysics Laboratory at University of Texas (UT) at Galveston as they pored over some global seismicity data in 1977.1
The Texas team realized that a 185-mile stretch of seacoast was involved—an area which had stayed still while areas of coastline to the northwest or to the east (into Central America) had suffered great quakes much more recently. Yet the Oaxaca area of Mexico’s coast sits atop a classic seismic bomb—a highly active subduction zone in which the floor of the Pacific Ocean (the Cocos plate at that point) is being shoved beneath the North American plate near its southern border with the Caribbean plate. Since great quakes had occurred many times in Oaxaca’s past, and since great quakes had occurred on either side of that 185-mile-wide sector in recent years, and further, since the rate of subduction for the plates could be assumed to be the same for the Oaxaca sector as well as the adjacent sectors, logic would dictate, they pointed out, that at some point in the not so distant future another great quake would take place under or near Oaxaca.
In other words, the Oaxaca sector was a “seismic gap,” an area along a fault or subduction zone which could be identified as due or overdue for an earthquake because of the rate of plate movement and the history of earthquakes on either side of the gap.
It was the validation of the theory, not the theory itself, that was new. Japanese seismologist A. Inamura had identified Tokyo as sitting on a seismic gap before 1923. Since continents were thought to be stationary at that time, his method involved a study of past earthquakes in the Tokyo area. He had concluded that if the adjacent coastal areas were periodically hit by major quakes, and if they had suffered such quakes more recently than the last one in Tokyo (as was the case), then Tokyo was overdue. Tokyo, he said, must be sitting on a seismic gap—a gap which would be considered filled only when an earthquake of substantial size occurred.
Sure enough, Tokyo was hit in 1923 by a devastating quake which touched off a ruinous fire storm and killed 140,000 Japanese. Five years later, in 1928, Inamura identified another seismic gap: the Tokaido-Nankaido region in southwestern Japan, which was subsequently “filled” by two more great quakes in 1944 and 1946—events which further verified that he was on to a new method that might just mark the threshold of earthquake forecasting.2
But in Oaxaca’s case, there was another factor. Not only had the region been devoid of major earthquakes, it had been devoid of all earthquakes large and small since 1973. Since subduction zone pressures never stop, sudden cessation of seismic activity in the vicinity of a subduction zone means something has really welded the plates together far below. Naturally that “weld,” or asperity, cannot last beyond a certain level of strain, so the day inevitably will come when the snag breaks, the plates lurch forward, massive amounts of previously stored tectonic energy are released as the strain dissipates, and the region is again lashed by seismic waves.
A developing theory of subduction zone quakes held that such seismic gaps went through two phases before the inevitable major earthquake. During a so-called alpha phase (which might start many decades after the last major earthquake), all activity large and small would cease. Seismographs in the area would register little more than large ocean swells hitting the beach, passing trucks, and occasional hard landings by airliners at nearby airports. The area, in other words, would be locked tight tectonically, and would be under great, increasing stress.
The end of the alpha phase and the beginning of what was called the beta phase would be heralded by small earthquakes resuming in the area. If those small quakes and swarms of quakes (many of them too tiny to be felt by the population) resulted from the progressive fracturing of rock preparing to break under increasing strain, then there would be only a certain preparation period before the thunderous impact of a large or great quake completed the cycle.
Oaxaca, said the Texans, seemed to be in a classic alpha phase. Therefore, they reasoned, at some point in the next decade the beta phase should begin, announced by a sudden onset of small or moderate earthquakes. According to the theory, once the beta phase had begun, a great quake could not be far behind.
Discovering what a beta phase pattern looked like in living black-and-white on a seismogram, and getting a finite feel for the length of time between the beta phase onset and the main event (a Ms 7.5 earthquake or greater), were problems the Texans were eager to address. Such answers take research money, so with renewed optimism that the 1977 earthquake hazard reduction act (which was moving toward passage in the Congress) might open the floodgates for increased research funding of any program that might advance the science toward the goal of earthquake prediction, the University of Texas researchers submitted a proposal to the USGS for a two-year study costing a quarter of a million dollars. The plan was to place a sophisticated network of seismographs around the Oaxaca area in hopes of “capturing” the inevitable beta phase (as well as the eventual main earthquake). The data they could record from such a network might go a long way toward identifying the telltale signs of a subduction zone earthquake about to occur. A quarter of a million seemed a small price to pay for possible clues to prediction of great quakes that could kill thousands of people and ruin billions of dollars of property (in the absence of extensive earthquake hazard reduction programs).
The USGS funding officer in Menlo Park, California, however, turned it down. The 1977 bill might send a wave of new funding into the USGS the following year; but this was fiscal 1977, and the funds just weren’t available yet.3
In the meantime, storm clouds of a sociological nature were gathering over the Mexican coast—and over Galveston. The UT, Galveston team’s conclusions had been interpreted inadvertently in Mexico as a “prediction,” altered from a long-range statement of probability into a forecast by a local psychic (quoted in the press), and twisted into a specific prediction by the members of a scientifically uninformed population (who had filled in the blanks on their own and arrived at a precise date the earthquake was “supposed” to occur). The entire affair had escalated within months to a near panic in Oaxaca, which culminated with a nonevent: The earthquake did not occur as “scheduled.”4
The Mexican authorities were, to say the least, less than impressed with the state of advancement in earthquake prediction—especially that aspect which permitted carefully hedged technical forecasts to balloon into unscientific predictions of date, time, place, and magnitude.5
Karen McNally and s
everal other junior and senior scientists at Caltech had followed the ideas, reports, papers, proposals, and local overreactions with great interest, looking at the same data the Texas team had digested and agreeing that the Texans’ conclusions were probably right. It did appear that Oaxaca was in an alpha phase, and it was quite possible it would emerge at any time into the beta phase. It would be a tragedy if the complex seismic waves large and small that would course through the Oaxaca region when that happened were not carefully recorded for years of future study.
It was with such knowledge, and against such a background, that Dr. McNally flew to Mexico City in August 1978 at the invitation of the Institute of Geophysics (a division of the Universidad Autónoma de Mexico) to deliver a lecture to the Mexican seismologists and geophysicists looking at the same question: Is something about to happen in Oaxaca that bears close watching?
It was somewhat stuffy in the small auditorium in Mexico City—a combination of the Mexican capital’s high altitude and the heat of a near-equatorial summer. Though she was enjoying the topic, it was with some relief that Karen McNally began wrapping up the talk, heading toward the question-and-answer period that would probably be the most productive part of the day. Those arrayed before her all were graduate students or practicing scientists, listening intently as she described what research American seismologists were pursuing in California, how it pertained to Mexico, and her view of the Oaxaca situation.
Speaking English (her Spanish was good but somewhat limited), Karen’s voice rose and fell over a surprising range of volume, a voice of cultured smoothness on a wave of expressive tones. The daughter of highly educated Danish parents who had settled in Central California’s San Joaquin Valley, her accent bore the hints of European pronunciation, rounding out a captivating verbal style that was holding the rapt attention of her English-speaking Mexican hosts and colleagues as she told them of her recent efforts to find identifiable patterns in the seismograms of recent earthquake swarms which had preceded great quakes in other parts of the world. To what extent were they precursors? Was there a characteristic “signature” of such seismic noise before a great quake? Was there, at least, a characteristic pattern in each particular locality before a great quake? Could they find, in other words, a specific pattern of small and moderate earthquakes which could be read like a fingerprint—the fingerprint of a killer subduction zone quake lurking just around the corner? Perhaps someday they could point to such a pattern and say, in effect, “This is the signature of an upcoming Alaskan quake,” or, “This is the unmistakable pattern of an impending Ms 8.0 Chilean quake.”
Dr. Karen McNally had been walking them verbally through a logical matrix of possible solutions when the fact that a small bell had rung somewhere in the auditorium gave way to the realization that someone had stuck his head in the side door, interrupting her.
“There’s been an earthquake,” he said simply, and in English, to a group of scientists who had felt nothing.
It was a rather startling sight to passing students, the gaggle of seismologists, geophysicists, and geologists following their guest lecturer and spilling out of the auditorium en masse (after having instantly decided to abandon the question-and-answer period for some more specific answers regarding the announced quake), headed at flank speed for the old, imposing Tacubayn Observatory adjacent to the university. The huge, ancient metal gates of the place creaked open slowly as the elderly record keeper let them in, conducting them at a slower, more respectful pace—like a monk leading the rabble into the inner sanctum of an ancient monastery. There were three small seismic networks in Mexico, two of them digital and modern, the third—and oldest—consisting of very basic seismographs whose inked needles on the revolving drums recorded the “bare bones” wave arrivals from a few sensors around Mexico City and a couple of outlying areas. It was to these elderly machines that the group went, fingers pointing to ink scratches on white paper, pencils flying and calculators clicking as they pinpointed the P and S wave arrivals, calculated the time difference in between, applied the mathematical ratio representing the different speeds of the two types of waves, and derived the distance. They were in Mexico City, which had been unaffected. The quake had to be on the coast. Looking at the north-south drum, the east-west drum, and the up-down drum (seismic waves are three-dimensional and, in classic seismic stations, are measured by three different seismographs, each representing one axis), Karen and company quickly calculated the direction from which the seismic undulations had come.
All the while Karen McNally knew that the Texas team had not been monitoring the Oaxaca coast in recent months and might not have seen any patterns of tiny quakes preceding this moderate one. The Texans had anticipated that Oaxaca might break its quiescence by showing how it had happened, historically, two times before. But could this be it? On one level it was far too perfect a solution—not to mention the timing. Here she was, vitally interested in just this subject and just this area. Was it possible that she could be standing here, giving a lecture on the very subject at the precise moment the region came alive? That seemed too fantastic. Too exciting. It was like finding you have five digits on a six-digit lotto ticket for several million dollars, and waiting with bated breath for the last digit. And, scientifically, this was indeed a type of lotto. Could that last number be hers?
Suddenly the calculations seemed to reach a consensus. It was a clear case of suspicions confirmed. The waves had come over three hundred kilometers, and from the south-southeast. They had come from a moderate earthquake in Oaxaca, and that could mean only one thing: The seismic quiescence had broken. The Oaxaca seismic gap had entered the beta phase, and a great quake couldn’t be too far behind.
Dr. Karen McNally was excited! They would have to move fast. But first, they would have to find some money. Karen knew that however limited were American funds, the funds her Mexican colleagues had to work with were dismal by comparison. The Mexican scientists did, however, have a cadre of volunteer scientists and students as well as vehicles and equipment, along with some limited research funds. Karen, on the other hand, had Jack Everendon, her project manager at the USGS, to whom she normally turned for research money, and she would have to convince him to help.
Quickly returning to Pasadena, Karen McNally assembled a team of seven Caltech scientists and graduate students, and talked Everendon into approving a transfer of $3,700 from an existing USGS-funded project to finance an expedition to the Oaxaca coast. It was very obvious that time was of the essence. Since the beta phase had begun, it could be anywhere from weeks up to eighteen months before the expected great quake occurred. If they were going to be able to find distinctive characteristics in the signature of seismic waves from the small earthquakes that would probably continue up to the main break, then they had to get an array of sensitive seismometers (as well as strong-motion versions which would not go off scale in large tremors) into the field rapidly. Mexico’s sparse, existing seismograph network was far too inadequate.
But $3,700? Everendon gave McNally approval for the transfer, not for new funds. But the project he was approving on an emergency basis essentially proposed to do for $3,700 what the seismologists and geophysicists at UT, Galveston had wanted to do for $250,000. Everendon seemed just as convinced Karen’s quest would fall short of finding a method of prediction, but for that price, it was worth a try.
By the first week in November 1978, the network of portable seismographs was in place around the state of Oaxaca—only seven stations scattered through the wild Mexican countryside, but a network of instruments sensitive enough to find the signature, if there was one. For approximately eight days the needles traced flat lines. Then a series of small tremors rattled the area, reaching only magnitude 3.2 Ms. After mid-November it quieted down again, and for nearly two weeks, the only thing happening seismologically at the surface was the arduous daily work schedule kept by the team in constantly visiting each machine, checking and resetting the time signals (using satellite clocks for pre
cision), changing paper, performing maintenance, and driving—constantly driving—between sites.
On November 28, a second swarm of tremors shuddered the needles, these rising to Ms 3.7, sweeping from one end of the gap landward and toward the center.
And eighteen hours later, as Karen McNally was landing at Los Angeles International Airport, the final snap occurred. With a magnitude of Ms 7.8, the great waves of seismic energy lashed out for nearly a minute, rocking and undulating the Mexican countryside (sparsely populated in that area), damaging rural towns, throwing down some stone walls, cracking adobe houses, and damaging some buildings as far away as Mexico City, 290 miles to the northwest. Had such a quake occurred underneath Los Angeles, the results would have been catastrophic. Because of its location, however, no one died.
When McNally’s team members looked at the data, it was incredible. They had trapped the earthquake exactly in the middle of their portable seismographic net. With strong-motion seismograph tracings coupled with the weeks of buildup seismic activity, they had amassed an unprecedented record of a great subduction zone earthquake, including the train of microquakes that would never have registered meaningfully on the main Mexican seismic networks.
As the team members prepared to wait out the aftershock sequence (and Dr. McNally returned at high speed from L.A.), they all were very aware of what success could mean. If standard signatures could be found—if typical foreshock patterns could yield reliable formulas for calculating how long a beta phase would last before a major or great quake—seismology could take existing long-term predictions based on the identification of seismic gaps, and turn them into viable intermediate-term predictions. In fact, with enough work and enough data and a good measure of cooperation on the part of nature, it just might be that foreshock patterns could also provide reliable short-term prediction.
On Shaky Ground Page 25