14 Introduction
or football, 5 but far less than bicycles, for risk of loss of
life. Similarly, the $20 billion worth of damage from the
Northridge earthquake, though enormous, is about 10% of the
annual loss due to automobile accidents. As a result, earthquakes pose an interesting challenge to society because they
cause infrequent, but occasionally major, fatalities and damage. Society seems better able to accept risks that are more
frequent but where individual events are less destructive. 6
Similar issues surface when society must decide the costs,
benefits, and appropriateness of various measures to reduce
earthquake risks. Conceptually, the issues are essentially those
faced in daily life. For example, a home security system costing
$200 per year makes sense if one anticipates losing $1000 in
property to a burglary about every five years ($200/year), but
not if this loss is likely only once every 25 years ($40/year).
However, the analysis is difficult, because the limited historical
record of earthquakes makes it hard to assess their recurrence
and potential damage.
Seismology is used in various ways to try to mitigate earthquake risks. Studies of past earthquakes are integrated with
other geophysical data to forecast the location and size of
future earthquakes. These estimates help engineers design
earthquake-resistant structures, and help engineers and public
authorities estimate and prepare for future damage by developing codes for earthquake-resistant construction. Seismology is
also used by the insurance industry to develop rates for earthquake insurance, which can reduce the financial losses due to
earthquakes and provide the resources for economic recovery
after a damaging earthquake. Rates can be based on factors including the nature of a structure, its location relative to active
faults, and soil conditions. Homeowners and businesses then
decide whether to purchase insurance, depending on their perceived risk and the fact that damages must exceed a deductible
amount (10–15% of the insured value) before the insurance
company pays. A complexity for the insurer is that, unlike
automobile accidents, whose occurrence is relatively uniform,
earthquakes or other natural disasters are rare but can produce
concentrated damage so large as to imperil the insurer’s ability
to pay claims. Approaches to this problem include limits on
how much a company will insure in a given area, the use of
reinsurance by which one insurance company insures another,
catastrophe bonds that spread the financial risk into the global
capital market, and government insurance programs.
1.2.2 Engineering seismology and earthquake engineering
Most earthquake-related deaths result from the collapse of
buildings, because people standing in an open field during a
large earthquake would just be knocked down. Thus it is often
stated that in general “earthquakes don’t kill people; buildings
5 These figures are for American football; in other countries soccer, termed football
there, is safer for players but more dangerous for spectators.
6 For example, although considerable attention is paid to aviation disasters and
safety, far more lives could be saved at far less cost by enforcing automobile seat belt
laws.
kill people.” As a result, proper construction is the primary
method used to reduce earthquake risks. This issue is addressed
by engineering seismology and earthquake engineering, disciplines at the interface between seismology and civil engineering. Their joint goal is to understand the earthquake ground
motions that can damage buildings and other critical structures, and to design structures to survive them or at least ensure
the safety of the inhabitants.
These studies focus on the strong ground motion near earthquakes that is large enough to do damage, rather than the much
smaller and often imperceptible ground motions used in many
other seismological applications. Two common measures are
used to characterize the ground motion at a site. One is the acceleration, or the second time derivative of the ground motion.
Accelerations are primarily responsible for building destruction. A house would be unharmed on a high-speed train going
along a straight track, where there is no acceleration. However,
during an earthquake the house will be shaken and could be
damaged if the accelerations were large enough. These issues
are investigated using seismometers called accelerometers that
can operate during violent shaking close to an earthquake but are
less sensitive to the smaller ground motion from distant earthquakes. The seismic hazard to a given area is often described
by numerical models that estimate how likely an area is to experience a certain acceleration in a given time. For example, the
hazard map in Fig. 1.2-3 predicts the maximum acceleration
expected at a 2% probability in the next 50 years, or at least
once during the next 2500 (50/0.02) years. These values are
given as a fraction of “g,” the acceleration of gravity (9.8 m/s 2 ).
A second way to characterize strong ground motion uses
intensity, a descriptive measure of the effects of shaking.
Table 1.2-4 shows values for the commonly used Modified
Mercalli intensity (MMI) scale, which uses roman numerals
ranging from I (generally unfelt) to XII (total destruction).
Intensity is not uniquely related to acceleration, which is a
numerical parameter that seismologists compute for an earthquake and engineers use to describe building effects. The table
shows an approximate correspondence between intensity and
acceleration, but this can vary. However, intensity has the
advantage that it is inferred from human accounts, and so can
be determined where no seismometer was present and for
earthquakes that occurred before the modern seismometer was
invented (about 1890). Although intensity values can be
imprecise (a fallen chimney can raise the value for a large area),
they are often the best information available about historic
earthquakes. For example, intensity data provide much of
what is known about the New Madrid earthquakes of 1811
and 1812 (Fig. 1.2-4). These large earthquakes are interesting
in that they occurred in the relatively stable continental interior
of the North American plate (Section 5.6). Historical accounts
show that houses fell down (intensity X) in the tiny Mississippi
river town of New Madrid, and several chimneys toppled
(intensity VII) near St Louis. Intensities can be used to infer
earthquake magnitudes, albeit with significant uncertainties.
These data have been used to infer the magnitude (about 7.2 ±
or football, 5 but far less than bicycles, for risk of loss of
life. Similarly, the $20 billion worth of damage from the
Northridge earthquake, though enormous, is about 10% of the
annual loss due to automobile accidents. As a result, earthquakes pose an interesting challenge to society because they
cause infrequent, but occasionally major, fatalities and damage. Society seems better able to accept risks that are more
frequent but where individual events are less destructive. 6
Similar issues surface when society must decide the costs,
benefits, and appropriateness of various measures to reduce
earthquake risks. Conceptually, the issues are essentially those
faced in daily life. For example, a home security system costing
$200 per year makes sense if one anticipates losing $1000 in
property to a burglary about every five years ($200/year), but
not if this loss is likely only once every 25 years ($40/year).
However, the analysis is difficult, because the limited historical
record of earthquakes makes it hard to assess their recurrence
and potential damage.
Seismology is used in various ways to try to mitigate earthquake risks. Studies of past earthquakes are integrated with
other geophysical data to forecast the location and size of
future earthquakes. These estimates help engineers design
earthquake-resistant structures, and help engineers and public
authorities estimate and prepare for future damage by developing codes for earthquake-resistant construction. Seismology is
also used by the insurance industry to develop rates for earthquake insurance, which can reduce the financial losses due to
earthquakes and provide the resources for economic recovery
after a damaging earthquake. Rates can be based on factors including the nature of a structure, its location relative to active
faults, and soil conditions. Homeowners and businesses then
decide whether to purchase insurance, depending on their perceived risk and the fact that damages must exceed a deductible
amount (10–15% of the insured value) before the insurance
company pays. A complexity for the insurer is that, unlike
automobile accidents, whose occurrence is relatively uniform,
earthquakes or other natural disasters are rare but can produce
concentrated damage so large as to imperil the insurer’s ability
to pay claims. Approaches to this problem include limits on
how much a company will insure in a given area, the use of
reinsurance by which one insurance company insures another,
catastrophe bonds that spread the financial risk into the global
capital market, and government insurance programs.
1.2.2 Engineering seismology and earthquake engineering
Most earthquake-related deaths result from the collapse of
buildings, because people standing in an open field during a
large earthquake would just be knocked down. Thus it is often
stated that in general “earthquakes don’t kill people; buildings
5 These figures are for American football; in other countries soccer, termed football
there, is safer for players but more dangerous for spectators.
6 For example, although considerable attention is paid to aviation disasters and
safety, far more lives could be saved at far less cost by enforcing automobile seat belt
laws.
kill people.” As a result, proper construction is the primary
method used to reduce earthquake risks. This issue is addressed
by engineering seismology and earthquake engineering, disciplines at the interface between seismology and civil engineering. Their joint goal is to understand the earthquake ground
motions that can damage buildings and other critical structures, and to design structures to survive them or at least ensure
the safety of the inhabitants.
These studies focus on the strong ground motion near earthquakes that is large enough to do damage, rather than the much
smaller and often imperceptible ground motions used in many
other seismological applications. Two common measures are
used to characterize the ground motion at a site. One is the acceleration, or the second time derivative of the ground motion.
Accelerations are primarily responsible for building destruction. A house would be unharmed on a high-speed train going
along a straight track, where there is no acceleration. However,
during an earthquake the house will be shaken and could be
damaged if the accelerations were large enough. These issues
are investigated using seismometers called accelerometers that
can operate during violent shaking close to an earthquake but are
less sensitive to the smaller ground motion from distant earthquakes. The seismic hazard to a given area is often described
by numerical models that estimate how likely an area is to experience a certain acceleration in a given time. For example, the
hazard map in Fig. 1.2-3 predicts the maximum acceleration
expected at a 2% probability in the next 50 years, or at least
once during the next 2500 (50/0.02) years. These values are
given as a fraction of “g,” the acceleration of gravity (9.8 m/s 2 ).
A second way to characterize strong ground motion uses
intensity, a descriptive measure of the effects of shaking.
Table 1.2-4 shows values for the commonly used Modified
Mercalli intensity (MMI) scale, which uses roman numerals
ranging from I (generally unfelt) to XII (total destruction).
Intensity is not uniquely related to acceleration, which is a
numerical parameter that seismologists compute for an earthquake and engineers use to describe building effects. The table
shows an approximate correspondence between intensity and
acceleration, but this can vary. However, intensity has the
advantage that it is inferred from human accounts, and so can
be determined where no seismometer was present and for
earthquakes that occurred before the modern seismometer was
invented (about 1890). Although intensity values can be
imprecise (a fallen chimney can raise the value for a large area),
they are often the best information available about historic
earthquakes. For example, intensity data provide much of
what is known about the New Madrid earthquakes of 1811
and 1812 (Fig. 1.2-4). These large earthquakes are interesting
in that they occurred in the relatively stable continental interior
of the North American plate (Section 5.6). Historical accounts
show that houses fell down (intensity X) in the tiny Mississippi
river town of New Madrid, and several chimneys toppled
(intensity VII) near St Louis. Intensities can be used to infer
earthquake magnitudes, albeit with significant uncertainties.
These data have been used to infer the magnitude (about 7.2 ±