Â鶹´«Ã½

IT WAS LATE afternoon on 17 October 1989, and millions of people in
San Francisco were beginning to think about going home after the day’s work.
Elsewhere, Americans were tuning their television sets into the third match
of the championship baseball series between the San Francisco Giants and
their neighbours across the bay, the Oakland Athletics – a game that was
going to take place at Candlestick Park in San Francisco.

A few minutes after 5 pm, a major earthquake cracked the San Andreas
fault in the Santa Cruz Mountains, 100 kilometres to the south. The shaking
lasted less than 15 seconds, but it resulted in 67 deaths, more than 3500
people injured, about 12 000 people homeless, and damage to property costing
almost $6 billion. There was also major disruption of transport, water,
gas and electricity supplies, as well as communications, in Santa Cruz and
the San Francisco Bay Area.

An awestruck television audience watched the reactions of the 62 000
fans at the Candlestick Park baseball game. The whole stadium began to rumble
and the concrete roof on the upper deck of the three-level stadium looked
as if it was shimmering. The crowd stood up but there was no panic, principally
because the shaking was sharp, but short. As the rumbling subsided there
was a round of applause and then the chant ‘Play ball’ – after all, earthquakes
are a part of life in California.

In the end, most of San Francisco, the Bay area and Northern California
came through remarkably unscathed. Of course this was not ‘The Big One’,
the expected quake which will be as big as the one in 1906. Loma Prieta
was a tremor of magnitude 7.1 on the Richter scale, and seismologists estimate
that the 1906 quake was 8.1; each whole number on this scale means 10 times
more shaking of the ground, and 30 times more energy released. So if the
Loma Prieta earthquake was a ‘pretty big one’, what lessons have geologists
and seismologists learned from its effects on local land and buildings?
The first lesson was that although the earthquake was anticipated, it was
not predicted to the minute by the network of seismic and other geophysical
monitoring equipment that laces California along the San Andreas fault.

California is subject to earthquakes because the state includes the
San Andreas fault, the boundary between two of the Earth’s tectonic plates.
The Pacific Ocean plate is moving inexorably northwest relative to the North
American continental plate. Over the past 30 million years, this movement
has averaged about 40 millimetres per year. But the fault tends to make
these movements intermittently; years of stability end when the fault suddenly
ruptures and moves a few metres. Geologists can calculate when the next
large event is likely to happen in the long term, over decades or centuries.
Each quake has to keep up with the average slip over geological time. The
bigger the quake, the longer the probable time between it and the next one
(‘Living on the fault line’, Â鶹´«Ã½, 6 January 1990).

On the basis of these estimates, the portion of the San Andreas fault
that ruptured during the Loma Prieta earthquake was identified as early
as 1981 as an area where another quake was likely.

A 1988 study by the US National Earthquake Prediction Evaluation Council
assigned to this segment the highest probability of producing an earthquake
of magnitude between 6.5 and 7 of any fault in California north of the Los
Angeles metropolitan area.

The Loma Prieta quake also added weight to the ‘seismic gap’ hypothesis,
with which researchers estimate the likelihood that part of a fault will
set off an earthquake. Seismic gaps are segments of a fault where there
has not been any significant movement for many years; these sections produce
fewer small earthquakes than the fault at each side. Instead of slipping,
the fault sticks and strain builds up in the crust of the Earth, storing
potential energy which will be released in a major earthquake in the future.
In 1981, Allan Lindh and William Ellsworth of the US Geological Survey identified
the southern Santa Cruz Mountains segment of the San Andreas fault as just
such a seismic gap.

The hypothesis was confirmed when the Loma Prieta earthquake and its
many aftershocks almost completely filled the Southern Santa Cruz Mountains
gap (This Week, 28 October 1989). One reason for a seismic gap is that the
two sides of the fault are sticking together at an asperity. Think of sliding
two pieces of wood past each other; if they have smooth edges, they slide
steadily. Jagged edges, with asperities, stick.

There is another asperity in the San Francisco Peninsula. Geologists
have not had time to assess what effect, if any, the Loma Prieta event has
had on the stress across the faults in this region. Nevertheless, in January
1990, the US National Earthquake Prediction Evaluation Council revised its
estimate for the probability of an earthquake of magnitude 6.5 or larger
within the next 30 years, from 50 per cent up to between 60 and 65 per cent.

Another unexpected aspect of the Loma Prieta quake was the lack of rupturing
at the surface, above the 40-kilometre-long break on the San Andreas fault
underground. On past experience, earthquakes of this size usually produce
anywhere between 1 and 2 metres of horizontal slip at the surface.

In the region close to the epicentre, directly above the start of the
rupture on the fault below, geologists hunting for evidence of what had
happened saw only cracks with small horizontal offsets, in the same sense
as the overall motion of the San Andreas fault: the western sides moved
northwest, and the eastern sides southeast. Other areas of cracking made
complicated patterns difficult to relate directly to the fault; their origin
remains something of a mystery.

Seismologists were disappointed that the Loma Prieta quake, while forecast,
supplied no obvious short-term warning signs, despite the network of geophysical
monitoring instruments around this part of the San Andreas fault. With hindsight,
the only harbinger was a marked increase in the ever-present small earthquakes
over the 16 months before the event. But these tremors diminished without
anything that geologists might have positively identified as a foreshock
of the October quake. Surveys looking for changes in the lie of the land,
and measurements of strain taken down boreholes showed only slight variations
at roughly the same time as the seismic activity increased. None of these
changes stood out above the background noise level, and nothing focused
interest around Loma Prieta, although there were no monitoring instruments
placed close to the place that was to become the epicentre.

The lack of warning signals was obviously a disappointing lesson from
the Loma Prieta quake. Long-term seismic activity has so far been one of
the best clues to future behaviour of faults in California. Seismologists
still have another test of this theory, and of the geophysical monitoring
system, around Parkfield, in southern California. This marks another seismic
gap, where there have been four earthquakes of about magnitude 6 this century,
the last in 1966. The area is quiet for 22 years or so, then there is a
quake. Scientists from the US Geological Survey predict another earthquake
before 1993.

If the Parkfield tremor does not materialise, seismologists may have
to look to different warning signals. Quite by chance, Antony Fraser-Smith,
and his colleagues from the electrical engineering department at Stanford
University, may have stumbled on an unexpected new tool for predicting earthquakes
– radio waves.

The Stanford team has been experimenting with very low frequency (VLF)
radio waves to communicate with submarines. Conventional radio waves carry
more data, but the VLF waves, with frequencies less than 30 000 hertz, pass
more easily through the Earth and the sea. Recently, the Stanford group
has been monitoring ultra low frequency (ULF) waves, with frequencies between
0.01 and 10 hertz, in an attempt to look at background radio noise from
the Sun that might affect VLF transmissions. To avoid interference from
electromagnetic noise in built-up areas, they installed the receiving antenna
for the survey at Corralitos, a small town 7 kilometres from the epicentre
of last October’s earthquake.

A signal surprise

When the Stanford group retrieved their records afterwards, they noticed
that the amplitude of the ULF signals had increased about three hours before
the quake. They immediately checked that this effect was real rather than
a quirk of their equipment. So far, they believe that it is genuine. Moreover,
about 12 days before the earthquake the background radio noise across all
frequencies shot up, then abruptly decreased on the day before the tremor.

Such an odd signal may have been due to electric currents generated
in rocks by stress before they fractured in the quake. This could happen
through the piezoelectric effect, the voltage that develops when crystals
are squeezed. Many of the minerals that make up rocks, quartz, for example
show such a voltage. Alternatively, changes in strain along the fault could
have altered the natural pattern of electric current in the Earth. There
may be a link between earthquakes and electromagnetic activity similar to
that which generate the ‘earthquake lights’ recorded during a series of
earthquakes at Matsushiro in Japan between 1965 and 1967.

The Parkfield prediction experiment would be a good test of the effectiveness
of the VLF tool; Fraser-Smith and his colleagues are considering setting
up their antennas there, to await the quake with the geologists and seismologists.

Not all the lessons from the Loma Prieta quake were news to geologists:
much of the damage resulting from the quake reinforced ideas that have been
known since 1906. One of the most important of these is the way in which
rock and sediment close to the surface transmit the seismic energy in different
ways, thereby influencing the degree of shaking felt during the quake. This
earthquake damaged buildings, roads and bridges in San Francisco and Oakland
about 100 kilometres from its epicentre; a magnitude 8.1 earthquake devastated
Mexico City on 19 September 1985 even though its epicentre was more than
350 kilometres away. Damage was severe in both cases because the areas stood
on young, fine-grained sediments, poorly consolidated and saturated with
water. Mexico City lies on the bed of an old lake, and mud and sand form
the land at the edges of San Francisco Bay. The coast of the bay has been
extensively modified; in most cases rubble and sand has been simply dumped
on top of the original mud.

Particular problems during the Loma Prieta quake came from liquefaction,
the transformation of wet, yet solid sand, saturated with water, into a
thick fluid. This causes damage because sand in this liquefied state has
virtually no strength; buildings above can subside, and land on even the
gentlest slope can shift and flow. This kind of movement, known as lateral
spreading, can devastate buildings, bridges, roads and river banks, as well
as damaging underground gas pipes and sewers. Other visible effects of this
spreading are open cracks in material above the liquefied sand, and sand
boils. These structures, like small volcanoes, arise when the sand erupts
through the ground surface, under the extra pressure.

A particularly important lesson from the Loma Prieta earthquake was
that most of the liquefaction that resulted in damage in the 1906 earthquake
happened in the same places during the 1989 event.

There was most liquefaction in areas that had been artificially filled,
or where wet sand or mud lay beneath broad tarmac pavements such as runways
at Oakland International Airport and Alameda Naval Air Station. Streets
and concrete slab foundations, such as in the toll plaza of the San Francisco
Bay Bridge fared equally badly. Accounts of the effects of the 1865, 1868,
and 1906 earthquakes in the Market and Mission districts of San Francisco
differ little from what has been written about these areas in October 1989.
The most intense (and most televised) damage from liquefaction occurred
in the Marina district, a community of family houses and flats, built mostly
with wooden frames. Structures like these normally resist earthquake shaking
when they stand on a foundation of solid rock, yet many of them were damaged
or destroyed.

Although building design may have contributed to the devastation, liquefaction
was widespread; sand boils erupted into basements, streets, yards and parks,
and cracks opened on roads and pavements. Most of the broken underground
pipes, which left about a thousand homes without gas or water after the
earthquake, were severed by with this flow.

The 1989 earthquake followed a pattern in part set by the quake of 1906;
the fine sand fill on which the Marina district was built was put there
after the 1906 earthquake; sand was dumped into the water and left to settle.
Debris from buildings destroyed in 1906 may also have been used as landfill.
The Marina district stands on an old lagoon that was the site of the 1915
International Exposition, held to celebrate San Francisco’s rejuvenation
after the great earthquake. On 17 October 1989, sand boils erupted carrying
up pieces of charred redwood, tar paper, and other buried debris from 1906.

Vibrations from aftershocks support the conclusion that damage in the
Marina district was made worse because the fill beneath deformed. Three
seismograms of the same aftershock, from separate sites, differ markedly.
An instrument standing on sandstone showed slight shaking, while seismograms
placed on natural dune sand and artificial fill recorded much more violent
movement. The San Francisco Gas and Light building, constructed in 1893
from brick, stands on the sandstone site. This is not the sort of structure
that fares well during strong shaking, yet it rode out the recent earthquake
and that of 1906 without a crack.

The seismometer placed on artificial fill, just 150 metres away, registered
much more vibration, leaving houses badly damaged after their foundations
failed; the north side of the street is now almost 0.6 metres lower than
the south side.

Significantly, there was no obvious ground damage in extensive landfills
along the central and southern shores of San Francisco Bay. Most of these
are dumped earth and rock, silt and clay emplaced under water, or solid
waste covered by compacted earth, all designed to modern engineering specifications
to resist liquefaction.

The greater shaking on softer sediments played its part in one of the
most serious collapses, at the Cypress Street viaduct of the Interstate
880 Freeway in Oakland. Almost 3 kilometres of the double-decker road, built
from reinforced concrete, collapsed onto traffic that was unusually light
perhaps because of the baseball championship. It killed 41 people.

Seismologists from Lamont-Doherty Geological Observatory in New York
State and from the US Geological Survey have found that the collapsed section
was built on top of landfill sitting on bay mud. The collapse happened because
of the design and construction of the joint between the lower deck and the
upper deck columns. The columns were not reinforced adequately at the points
where they supported the decks. Plans to strengthen the road, prepared some
years ago, were not acted on because of a lack of funds.

The principal lesson for the future from the Loma Prieta quake was that
the pattern of damage was similar to that witnessed in 1906: the geology
plays an important part in determining where the fault ruptures are likely
to be, how hard the ground shakes, where landslides occur, and where the
ground will sink, spread or crack.

Good science and engineering alone are not enough to reduce the hazards
likely from earthquakes. It requires cooperation between a well-informed
and well-prepared public and government at all levels. Those living and
working in earthquake country should press for better construction practice,
improved building codes, and modification of structures that are dangerous.
We can either rigorously apply the lessons learned from the Loma Prieta
earthquake or be condemned to relearn them next time.

* * *

Maps that help to spot where trouble lies in store

THE LOMA Prieta earthquake provided a direct test of the techniques
of seismic zonation, which have been used to identify the potential for
hazards from earthquakes in different areas. The likelihood of large quakes
on the various faults, the influence of local geological conditions on the
amplitude of ground shaking and therefore the extent of damage expected
all contribute to mapping out the hazards.

Researchers make isoseismal maps, defining areas that experienced equal
shaking during an earthquake. The maps record intensity in terms of the
Modified Mercalli scale (MM), which takes into account how far a particular
site is from the source of the earthquake, the size of the quake and whether
the site has solid rock or softer, unconsolidated sediments underneath.
On a scale of I to XII, it gives a measure of how strong a quake feels at
different places, and therefore is useful in predicting where buildings
and structures will be at risk in future.

Qualitative data comes from the where there have been landslides and
damage to trees, buildings, bridges and the ground. Eyewitness accounts
of what happened in the quake are especially important. Quantitative records
of the horizontal and vertical displacements of the ground during earthquakes
come from a type of low-sensitivity seismograph, known as an accelerograph.

In the October quake, seismologists collected records of ground shaking
at over 170 sites within 200 kilometres of the epicentre. These instruments
are maintained by the California Division of Mines and Geology and the US
Geological Survey, and funded by a tax levied on permits required for new
buildings.

The intensity of the Loma Prieta earthquake was rated as MM VIII around
the epicentre, although isolated sites in the cities of San Francisco and
Oakland, more than 100 kilometres to the north, experienced intensity levels
of MM IX. Close to the epicentre, where geologists inferred that there had
been extremely hard shaking, the maximum acceleration recorded was 64 per
cent of gravity. Maxi mum accelerations recorded elsewhere depended on the
type of rocks beneath. For example, in Oakland, 100 kilometres away, two
of the three highest values, 26 per cent and 29 per cent of gravity, were
recorded on top of mud. The lowest values came from sites on solid rock.

In 1975, Roger Borcherdt and his colleagues at the US Geological Survey
made maps of the San Francisco Bay area, to show the potential for surface
faulting, ground shaking, liquefaction, landslides and flooding.

This information has already been used in developing land-use policies
across California. Some of it has been incorporated in public safety legislation.
The Alquist-Priolo Act of 1972 requires that a zone of special studies be
identified along the traces of known active faults that are capable of causing
earthquakes.

Unfortunately, much of the growth of San Francisco around the bay took
place before this act came into force. Already, there are houses, industry
and public service buildings on active traces of the Hayward Fault and,
to a lesser extent, on the San Andreas Fault. Development has also spread
to areas of landfill on top of mud from the bay, to other areas with high
potentials for liquefaction, and to slopes subject to landslides. Many buildings
within this region predate the modern building codes which require designs
that resist earthquakes.

The most densely-populated part of the San Francisco Bay area is vulnerable
to large earthquakes on either the San Andreas or Hayward Faults. Maps of
seismic risk for the area combine the damage recorded during the 1906 earthquake
with recent geological mapping and quantitative measurements of ground motion.

To help to show the usefulness of such maps for deciding how best to
use land, Earl Brabb and his colleagues from the US Geological Survey have
started a project in San Mateo County, south of San Francisco, to produce
a folio of maps at a scale of 1:62 500. These will show the potential for
surface faulting, ground shaking, liquefaction, landslides and structural
damage across the region.

* * *

Problems with planning for America’s quakes

IN 1977, the US Congress created the National Earthquake Hazard Reduction
Programme, with the aim of decreasing the death and destruction that come
from earthquakes. Every two years, Congress reassesses and reauthorises
the programme. And, this year, in the wake of the Loma Prieta earthquake,
the customary debate has acquired more force.

Four agencies carry the principal responsibility within the programme:
the Federal Emergency Management Agency, the US Geological Survey, the National
Science Foundation, and the National Institute of Standards and Technology.
FEMA’s role is that of coordinator, the NSF and USGS are responsible for
basic research while NIST does more applied research and establishes practical
standards.

Riley Chung is director of a board of the National Research Council
which is reviewing earthquake engineering research within this programme.
He says that FEMA has not been doing its job. As coordinator, FEMA is supposed
to understand what the other agencies do and to facilitate the transfer
of technology, for example from engineering theory to building practice.
Chung says that the Agency appointed the wrong staff, bringing in people
from their flood programme.

FEMA, though, is not completely to blame. Congress gave it the coordinating
role without the power to review research and say what money should be spent
on which projects. Instead, the NSF, USGS and NIST make their own decisions.

This issue has been brought up repeatedly, but Congress has taken no
steps to rectify the situation, says Chung. In their report, due out in
early summer, the NRC will recommend that Congress appoint some group with
the power to review what all of the agencies do, leaving the day-to-day
coordination to FEMA.

Chung also criticises Congress itself for lack of coordination. Each
of the principal agencies in the earthquake hazard reduction programme is
responsible to one or more congressional subcommittees. ‘Congress lacks
the coordination to look at the programme; it would be better if one subcommittee
had the responsibility.’

FEMA’s ineffectiveness as a coordinator has contributed to a gap between
scientific knowledge, such as the severity of shaking expected from a quake,
and practical application of that research to, for example, vulnerable structures.
But shortage of money is also to blame, according to Christopher Arnold,
president of the company Building Systems Development and one of the NRC’s
committee of experts on earthquake engineering research. He says: ‘If a
practising engineer receives an academic paper, it’s not much use. The engineer
doesn’t know whether it’s right or not. We have the fundamental knowledge,
but the resources to use them are not available.’

If more effort had been made to turn basic research into building codes,
Chung believes that the $20 billion to $30 billion costs in damage to property
resulting from the Loma Prieta quake could have been reduced.

Besides closing the gap between research and its applications, Chung
says that the hazards reduction programme needs more engineering research.

In testimony to the Senate subcommittee on science, technology and space,
Robert Kuntz, president of the California Engineering Foundation, said too
that money was needed to tell engineers about new ideas. ‘It seems . . .
that funds are not available to educate the practitioners’, said Kuntz.

Another issue that Loma Prieta highlights is that earthquake engineering
research needs to look specifically at ways to strengthen old buildings
and protect sewage, gas, electricity and water supplies. Most of the houses
that collapsed last October, says Arnold, had not been reinforced. And,
although the new high rise buildings had only minor damage, neither Arnold
nor Chung are complacent.

Arnold says: ‘Loma Prieta showed us the social and economic impacts
of a moderate earthquake. If anything, it was an exercise with live ammunition.
It was not really what we are worried about.’ To accomplish what is needed,
Arnold believes that spending on earthquake engineering research should
increase two- or threefold.

Henry Spall is a geophysicist with the US Geological Survey at their
National Centre in Reston, Virginia. He edits the Survey’s magazine Earthquakes
and Volcanoes.

Further Reading Lessons learned from the Loma Prieta, California, earthquake
of October 17, 1989, edited by George Plafker and John P. Galloway, 1989,
US Geological Survey Circular 1045, available free from US Geological Survey
Books and Open-File Reports, Federal Center, Box 25425, Denver, Colorado
80225, USA; The Loma Prieta, California, earthquake: an anticipated event,
US Geological Survey staff, 1990, Science, vol. 247, p 286