INTERVIEW

ALLAN LINDH

*
Earthquake depths are less accessible than the moons of Jupiter, says a top
seismologist, who predicts a one-in-five chance of a Big One in the San
Francisco Bay Area in the next five years

In 1983 Allan Lindh forecast the Loma Prieta earth quake. He did not, of
course, say it would occur on October 17, 1989-as it did nor even that it
was due before the turn of the century. What he said was he saw a 5 to 90
percent chance of a major quake in that California segment of the San
Andreas Fault within the next 30 years.

Lindh, a top seismologist with the U.S. Geological Survey (USGS), was not
alone in his suspicions of that particular segment of the San Andreas Fault. In
the 16 months before the Loma Prieta earthquake, the USGS issued two warnings of
increased risk in that area. Afterward most agreed that "the earthquake Al Lindh
predicted took place," as colleague Wayne Thatcher put it.

In the fall of 1990 Lindh did not predict a 50 percent chance of an earthquake
along the New Madrid Fault on December 3. In fact, he thought it highly
implausible that an earthquake would occur in that south-central area at that
time. When December 3 came and went and there was no quake, Lindh took
this fact to be a "direct sign of the existence of God, who chose to
signal that She was on the side of the real seismologists."

Earthquake prediction is still a novelty in the world of respectable
science. Until a few decades ago the mere idea of a scientific basis for
forecasting was a laugh. "Only fools, charlatans, and liars predict
earthquakes," said Charles Richter. But growing knowledge of plate
tectonics in the Seventies gave some rationale to the sudden ruptures
of the earth. We now understand that an active fault zone is the result
of two continent-sized plates grinding by each other. In some places
movement is gradual; in others rocks lock together until the pressure
becomes overwhelming, whereupon an earthquake snaps them past each other.

So earthquake prediction should be simple. All you have to know is
how fast the plates are moving and how much plate divergence leads to
a quake, and you've got it. But even with adequate earthquake history
and good measurements (very tricky), the result is still approximate. The
current idea of a prediction, or "forecast," is to posit along the lines
of a 30 percent chance in the next 30 years.

Lindh is trying to narrow the span. A leading force in establishing the
USGS's Parkfield Earthquake Experiment, today he is lobbying to wire the
San Francisco Bay Area with earth movement analyzing instruments. Calling
itself "the earthquake capital of the world," Parkfield (population 34)
in central California averages a medium-sized quake about every 22
years. In 1985 the federal government issued its first official earthquake
prediction-a 90 percent chance of a magnitude six by 1993-based on
research at Parkfield. The last Parkfield quake was in 1966, and the
next is overdue. Meanwhile several million dollars worth of measuring
devices have been sunk into the Parkfield earth in the hope that after
the quake to come, Monday-morning quarterbacks will be able to sort out the
signals unique to imminent earthquakes.

Lindh, forty-seven, has spent his career with the USGS in Menlo Park,
California. Having received his Ph.D. at nearby Stanford, he lives with
his family about half a mile from the San Andreas Fault. A beeper on
his belt informs him of most significant quakes in Northern California.
Lindh submits that his career as a seismologist began with cleaning
ducks at age five. The route from ducks to seismology was not direct.
After dropping out of college during the Cuban Missile Crisis, he spent
the next decade driving a garbage truck, working in a Canadian oil
field, and arguing with his draft board. When he finally arrived at the
University of California at Santa Cruz, Lindh studied geophysics. Part of
the appeal of seismology, he admits, was that so little was known
about it then a young man in a hurry could pick it up pretty fast.

San Francisco writer Esther Wanning toured Parkfield with Lindh-and his
office, where newspapers, unfiled reports, and rolled-up charts made a
landfill that raised the floor level several feet. Like many people too
busy to organize, Lindh demonstrated an uncanny ability to dive through
the crust and produce the article under discussion. A path led to the
desk, where a PC with ties to earthquakes worldwide reigned. On
Wanning's second visit, Lindh, in deference to an out-of-kilter back,
stretched out with his head nearly under his desk and the tape
recorder resting on his stomach.

Twice beepers sounded and seismologists from all over the building
converged on a set of seismometers recording quakes, As neither
disturbance proved alarming, the crowd quickly disbanded. But every
little quake adds to the big picture. The USGS center in Menlo Park
issues a weekly seismicity report for Northern California-invented,
nurtured, and written by Lindh. The report, which describes the past
week's rumbles and the changing prognosis for future quakes, is widely
circulated and nervously watched by the media. For Al Lindh, earthquake
prediction is anything but academic.

Omni: When climatologist Then Browning predicted the quake on the New
Madrid Fault, his theory correlated earthquakes with high tidal forces.
Were you nervous he might be right?

Lindh: No. I knew the odds were no different for December third than
for any other day, when they are probably about one in a hundred
thousand. In fact, I've never seen such a low level of seismicity
around the world as there was during this five-day period. I doubt if
anyone will ever take seriously again the question of tidal triggering
of earthquakes. Careful examination of earthquake data over decades shows
that there isn't anything to it. Theoretically, it's plausible. The sun
and the moon do distort the earth, and one would expect that tides
would trigger a quake some of the time. The mystery is that they do not.

Omni: Browning supposedly had predicted previous earthquakes.

Lindh: If you predict enough events, some of them are bound to be
right. Then if you selectively recall the ones that correspond with
something, you can claim with a straight face that you can predict
earthquakes.

Omni: Why was he taken seriously?

Lindh: Because the low-rent journalists who picked up the story wanted
headlines and filled them with irresponsible distortions. The good
science reporters adamantly avoided the topic.

Omni: Where were you when the quake of October 17, 1989, struck?

Lindh: In my truck with my feet hanging out, watching my son's
soccer game. The field was in the middle of a primeval forest, and
the trees swayed back and forth along with my truck. I enjoyed it. I
had no idea what it was. It seemed too small to be the one we
anticipated on the Hayward Fault, and too big for Loma Prieta. With so
many faults close at hand there was no way of knowing immediately.

Omni: But you'd predicted it.

Lindh: That's putting it too strongly. Prediction is not a
black-and-white issue, but a long continuum of grays from long-term
projections to short-term stuff. If you rate predictions on a scale of
one to ten, we got a three for Loma Prieta. In 1982 we specified an
earthquake on the Loma Prieta segment in 1988-plus or minus seven
years. But we became more wishy-washy in the intervening years. We hit
the spot right, but the earthquake slipped more, went deeper. So our
projected 6.5 turned out to be a 7.1 [on the Richter scale].

Earthquake depths are less accessible than the moons of Jupiter. With
action taking place ten or twelve kilometers down, the strain signals
are hard to measure because they decay quickly. Even for large events,
the surface manifestation is very small. Quakes are the archetypal
nonlinear process. The earth takes hundreds of years storing vast
amounts of energy and then gives it all back in seconds. You don't
need chaos theory to tell you that if you don't know much about the input
parameters, you won't know much about extreme nonlinear behavior. That's
the downside. The up side is that foreshocks, occurring within the
prior twenty-four hours near the epicenter, tell us there's action before
a quake. It's like when you bend a stick; you hear crackling before
it snaps. If it weren't for foreshocks, prediction might still be
something of a bad joke. Another problem is that each segment probably
has different characteristic symptoms.

Omni: Then what do you go by?

Lindh: The starting point is to divide the amount the earth moved in
the last quake by the slip rate. That's very approximate, and you can
only expect to get within about ten percent of the recurrence time. We
can measure ground movement with lasers and satellites to find how fast
the plates are moving, but we only have a vague idea what the critical
point for rupture is. We knew the Loma Prieta segment was dangerous
because the southward-propagating rupture of the San Andreas ran out of
steam there in 1906. The northern end of the fault slipped about sixteen
feet in 1906, whereas the Loma Prieta segment slipped only four and a
half to seven feet. The 1989 quake was making up some of the
shortfall. We don't expect another quake on the northern end for a hundred
years.

Omni: Why have you turned your attention to the Hayward Fault?

Lindh: The Hayward Fault runs right through old urban areas in Oakland
and Berkeley and is very close to San Francisco, so a big earthquake will
be a monumental tragedy-like war, with caskets lining the roads and
gymnasiums stacked with body bags, The fault produced big quakes in the
last century, but we don't know on exactly which segments.

Other seismologists and I have floated a proposal for installing
instrument clusters along two or three sensitive fault segments in the Bay
Area. Next I'd wire the San Bernardino area in Southern California. The
most dangerous part of the southern San Andreas, in the Coachella Valley,
is quite a ways from Los Angeles. The threat to L.A. depends on
whether three segments down there go as one. If they do, L.A. gets hit
hard. But the San Bernardino Mountains are a tremendous unknown in the
fault. Whether the San Andreas can rupture through those mountains in
one great quake is an open question.

Omni: What are the odds for a major earthquake in the Bay Area?

Lindh: Our latest estimates are sixty-seven percent for a major quake
in the next thirty years. That's twenty-eight and twenty-three percent for
the northern and southern segments of the Hayward Fault, and twenty-three
percent for the San Francisco peninsula segment of the San Andreas. The
odds for the peninsula segment went up after Loma Prieta, as we assume
Loma Prieta put more pressure on the segment north of it. It's
reasonable to speak of a one-in-five chance for a major quake in this region
in the next five years. Partially because in the past, earthquakes on the
San Andreas and Hayward faults seem to have come in pairs. Possibly,
though, that's just coincidence. We do know that recently we've entered
into a seismically active period.

Omni: It gives me chills.

Lindh: Me, too. In a big quake on the Hayward Fault, we'll see
ground motion maybe five times as great as Loma Prieta, and it'll go on
much longer, The nineteenth-century quakes did a lot of damage, and
there was wilderness where there are now major cities with lots of
people packed into old buildings. And because of the amount of
landfill, there's great liquefaction potential along the waterfront. A state
of California scenario for a magnitude 7.5 on the Hayward estimates
up to four thousand dead and injuries in the tens of thousands. A lot
of the energy from a quake on the Hayward Fault will end up in the
soil structure the freeways and bridges are built on. These cities will
suffer from both ends of the fault.

Omni: Why are there so many other faults out there?

Lindh: Because the earth has a crust like raisin pudding. It's got
four billion years of history, and it's very complicated in the upper
hundred kilometers. In Iowa the crust is a pretty solid structure, but in
California it's been holy hell for the last hundred million years with
the plates moving apart and smashing together. The fault would like to
be straight and clean from Mexico to Oregon, but things like the
Sierras get in the way, so it zigs and zags. Earthquakes take the path
of least resistance through the raisin pudding.

Omni: What are the odds of an earthquake east of the Rockies?

Lindh: One magnitude seven per century isn't a bad guess. But nobody
knows where or when. Florida doesn't seem to have many, and some of
the cold, stable Midwestern states don't produce many dots on the map.
Heaven only knows what's providing the strain energy driving Eastern
quakes. Apparently they occur on what's called failed rifts, places where at
some time a continent started to split, then stopped and left a defect.
There've been only two major quakes recorded east of the Rockies:
Charleston, South Carolina, in 1886, and New Madrid, in 1811. They were
both felt over most of the East because the East transmits seismic
energy very well. The New Madrid quake rang church bells in Boston.

Omni: What about an earthquake in New York City?

Lindh: A large quake is unlikely, and the unlikely event would be at most
a magnitude seven. But should such a thing happen, it would be too
terrible to think about. The Eastern building codes have paid little
attention to earthquakes; if these buildings had been on the West Coast
they'd have fallen down long ago. So in a magnitude seven earthquake,
tens or hundreds of thousands of people could die. New York City is
capable of the kind of disaster that occurred in Iran [in 1990]. But
you'd have to get an extremely improbable event in exactly the wrong
place.

We're miserably failing to give seismic hazards in the East the study
they deserve. It's a scandal. Over the past twenty years we've started to
build a fragmentary picture. But that work was driven almost entirely by
the rush to build nuclear power plants. Now that they're unfashionable,
there's little funding. This is grossly irresponsible. You can't wait
until 2050 to record the microearthquake that occurred in Ohio in 1991.
Our children and grandchildren will not have the data needed to make
hard decisions about where to put critical facilities, and they'll end
up making the same stupid decisions we do. The pity of it is that
continuing the existing network of seismometers would be such a cheap
and simple thing to do. You could do it for the kind of money that it
takes to water the flowers around the headquarters of nuclear power
plants.

Omni: In what part of the earth's crust do earthquakes occur?

Lindh: It's the upper twenty kilometers, in the cold, brittle part. Below
that, it's so hot there can't be the sudden slip that causes quakes.
It's probably more like hot taffy, and the movement goes on all the
time. We don't know how the brittle surface faults are connected to
what's going on down below. Is it pushing or pulling or is there a complex
interaction? But if I had a series of pictures of earthquakes over the
last few thousand years, I think the answers would be self-evident.

Omni: Can't you discover earthquake history by digging trenches?

Lindh: Kerry Sieh of Caltech and a few others have dug trenches in
carefully selected places and studied the upper ten to twenty feet of
sediments. Sediments adjacent to faults are disturbed every time there's a
big quake, and in lucky circumstances the record is preserved. Then more
sediments are laid on top. What you get is like a very complicated tape
recording of what happened. But sometimes sediments wash away, little
organisms stir them up, or the sedimentation is not the right type.
There are a million reasons why trenching might not work. Still, some
very smart young geologists are out there year after year doing the
dirtiest, most disgusting work a geologist has to do.

Geologists like to walk around on granite in the Sierras. They don't like to
set scaffolding and shoring down inside water-filled trenches where every
now and then a geologist is killed. Trenching is a dreadful business, but
it provides a unique picture. In one place Kerry Sieh has gone back
two thousand years and can now tell when an earthquake occurred to within
fifty to a hundred years. His work mainly has given us a prediction
for a sixty percent chance of a quake in Southern California in the next
thirty years. Trenching is a gamble, though cheap. With four or five
good people working for four or five years, we might be able to answer
the critical questions about the Hayward Fault. It takes a funny group of
people and hard work to make a run at earthquake prediction. Right now at
the USGS we still have the best-maybe the second- or third-best-bunch in
the world working on prediction. But if four thousand people die along
the Hayward Fault tomorrow, many of us will feel personally responsible.

Omni: What good will prediction do us? You can't stop an earthquake.

Lindh: As many as one quarter of the homes in the Bay Area are not
bolted to their foundations, which means twenty-five percent of the people
don't understand their life savings-usually the equity in their homes-can
disappear in seconds. Their negligence is criminal. There's a lot of
sloppiness in implementing building codes. Often buildings don't have any
bracing to support the beams or joists underneath. Persuading people that
there's a concrete reason for taking care of these little things is about our
most important function.

Omni: Do you think Bay Area chambers of commerce prefer not to
emphasize the coming quake?

Lindh: Well, too bad. People squawk about the money, but they forget, or
never heard, that the San Andreas is the single greatest economic boon to
California. The movement of the plates created the fault, but also the
harbors, agricultural valleys, mountains that store snow so you can
irrigate the desert lands, the folds that produce gas and oil, the
coastline and everything that makes California such a wonderful place.
How many tourists would come here if it looked like central Nevada?

Omni: Why do some people close to the fault have so little damage, and
others farther away have a great deal?

Lindh: Part is serendipity. Earthquakes don't send energy out uniformly
in all directions. In the Loma Prieta quake some really good houses
above Los Gatos were shattered. Those people were just unlucky. The
earthquake appears to have focused a lot of energy on that sharp-pointed
ridge. They may have experienced the highest intensities anyone in the
country has ever known. Houses leapt off their foundations, people were
thrown into walls and through windows, and refrigerators flew back and
forth across kitchens, Nonetheless no one died there because the houses
were well built.

Omni: What have you learned from the Parkfield Earthquake Experiment?

Lindh: We're trying to dope out how the long-term strain accumulation
and release cycles work, interact, and when they'll next reach the failure
point. But since we haven't had the earthquake yet, we haven't learned
how to predict them. We see strain building up, little creep events, a
shift in seismicity.

We're looking for any gradual failure. Before a fault slips fast-as in the
earthquake-it will have to slip slowly. We're looking for foreshocks or
anything on the strain meters that might go along with them. There's
evidence that suggests there's some action in the hours and minutes
before. Slip is what we're looking for. That's why we're trying to
build good strain meters and put them down in holes where they get
away from the earth's surface movement.

The important thing is to measure the deformation of the earth; the
earthquake machine down at depth is bending the rocks all the time. We
measure this deformation at Parkfield with a two-colored laser, measuring
movement of the plates millimeter by millimeter. Day in and day out we
see plate tectonics in action, Also people sometimes see a decrease in
background activity before earthquakes. Then the foreshocks, if there are
any, come out of the blue. We don't know if this change we've seen in
the last few years means anything,

Omni: Three hours before the Loma Prieta quake, the intensity of
electromagnetic waves skyrocketed.

Lindh: If we get an electromagnetic signal before the Parkfield quake like
that of Loma Prieta, the world of prediction will change very much. It
won't solve the whole problem. But you'd have more hope that when you
got down to the last few days, you'd be able to give people something
concrete.

Omni: What about the contention that animals sense coming earthquakes?

Lindh: The evidence is less than persuasive. Now, it may be that
water-level changes before quakes have driven snakes and rats out of their
holes. But if strain changes are so gross that you can see them in
shallow water, then we could put in instruments that would be cheaper
and more reliable than animals. But since the stories about
electromagnetic waves came up, it's been running through my head that
perhaps some animals some of the time-and even people-directly sense
changes in electromagnetic radiation. If true, and if there are big
electromagnetic frequency changes before quakes, it will reopen the
question of animal behavior before earthquakes.

Omni: How do the citizens of Parkfield feel about their coming quake?

Lindh: They rather enjoy the notoriety.

Omni: Are they afraid?

Lindh: Nah, If you're a cattle rancher, you've got more worrisome
things on your mind. Nobody's ever died in one at Parkfield. You'd have
to be very unlucky to be killed.

Omni: Will you ever be able to say, "Three days hence there will be
an earthquake in such and such a place"?

Lindh: I don't worry where the process will end up. We're decades
from that kind of capability.

Omni: We the public think you're going to tell us so we can plan for the
day.

Lindh: You're wrong. We've misled you, Moreover, you're not ready for
it. If we could, you wouldn't know what to do.

Omni: Well, tell us.

Lindh: No, we don't have any idea what you should do if told when
quake day was. We're all in this together.

But even if I thought there was only a one-in-a-thousand chance of
really predicting earthquakes, I'd still try for a way of understanding the
problem. Besides, earthquake research has led to seismic engineering and
building codes that have made quakes much less dangerous. And there's
our responsibility to the rest of the world, Roger Bilham at the
University of Colorado has pointed out there's a dreadful tendency for
the supercities that are growing up-especially in the Second and Third
worlds to be located in tectonically active areas. In the future we may see
millions, rather than thousands, die in earthquakes and the accompanying
fires. Our effort to understand earthquakes could be a greater
contribution to people in the Third World than the millions we send in
relief after a catastrophe.

Omni: But is it feasible to have some sort of real-time
thirty-to-sixty-second warning system?

Lindh: Maybe. Because the shear waves that carry most of the energy
travel more slowly than radio waves, we could detect an earthquake and
radio ahead to warn people at a distance from the epicenter of what's
coming, The instruments are the same as those for prediction. When you
build the prediction network, you build into it the capacity to respond in a
few seconds.

In the Midwest people buy radios that kick in with emergency tornado
warnings. For quakes, we'd provide the signal, and electronics stores
could sell little receivers. Within seconds we could provide an estimate
of how big the ground motion was going to be and how long it would go
on. People could hook computers up to this system to park discs. It'd
warn people in chemical plants and refineries and probably be used to
close some valves automatically-and give people time to run when a
chemical vat ruptures. The guy working under his car would beat it, too.

Omni: How did the Native Americans deal with earthquakes?

Lindh: The Kwakiutl, a Vancouver Island tribe, apparently built
quake-safe lodges, They also had quake dances that were very elaborate,
complete with simulations of shaking people and buildings. I think those
dances may have been a rough-and-ready hazard reduction practice. After
all, the real trick isn't how to build safe buildings, it's remembering to
do so, People know damn well how to build, but they forget what they
know. The Armenians have two thousand years of recorded history and yet
in the Sixties and Seventies they went right ahead and put up
buildings that killed as many as fifty thousand of them in 1988. The
Kwakiutl appear to have built good buildings.

The short attention span of humans is a fundamental problem in dealing
with long-term hazards. When transmitting knowledge from generation to
generation, you need not just the knowledge, but the emotional content
so that it gets acted upon. We're great at information nowadays, not so
great at passing along emotional content.