4 These terms are not the same as compressional and shear waves; as often occurs in
science, words have multiple meanings.
5 In discussing analogous issues Sarewitz and Pielke (2000) note than even after billions of dollars spent on climate research, a senior scientist observes, “This may come
as a shock to many people who assume that we do know adequately what’s going
on with the climate, but we don’t,” and the National Academy of Sciences states that
deficiencies in our understanding “place serious limitations on the confidence” of
climate modeling results.
surface, so geological and geodetic observations also show the
motion that occurs in earthquakes. In less accessible areas
seismological observations provide most of the data used to
identify the boundary along which motion occurs and to demonstrate its nature. This is the case for most plate boundaries,
which occur in the oceans, beneath several kilometers of water.
Similarly, in subduction zones, where lithospheric plates
descend deep into the mantle and earthquakes can occur to
depths of 660 km, direct observations are not possible, but
analyses of seismograms reveal the motions and give insight
into their tectonic causes.
1.1.2 Models in seismology
As summarized in the previous section, seismology provides a
great deal of information about seismic sources, the structure
of the earth, and the relation of earthquakes to the tectonic processes that produce them. Even so, we will see that there are
major limitations on what the present seismological observations and other data tell us. For example, although we have
good models of seismic velocity in the earth, we know much
less about the composition of the earth and have only general
ideas about the deep physical processes, such as convection,
thought to be taking place. Similarly, although seismology provides a great deal of detail about the slip that occurs during
an earthquake, we still have only general ideas about how
earthquakes are related to tectonics, little understanding of the
actual faulting process, no ability to predict earthquakes on
time scales shorter than a hundred years, and only rudimentary
methods to estimate earthquake hazards. This situation is
typical of the earth sciences, 5 largely because of the complexity
of the processes being studied and the limits of our observations. Our best response seems to be to show humility in face of
the complexity of nature, recognize what we presently know
occurred is inferred from the three-dimensional pattern of radiated seismic waves. Figure 1.1-7 illustrates the method used for
an earthquake in which the material on one side of a vertically
dipping fault moves horizontally with respect to that on the
other side. This motion generates seismic waves that propagate
away in all directions. In some directions the ground first
moves away from the source (toward a seismic station),
whereas in other directions the ground first moves toward the
source (away from a receiver). The seismograms thus differ
between stations. In the “toward” (called compressional)
quadrants the first ground motion recorded is toward the receiver, whereas in the “away” (called dilatational) quadrants
the first ground motion is away from the receiver. Because the
seismic waves go down from the source, turn, and arrive at a
distant seismographic station from below, the first motion
is upward in a compressional quadrant and downward in a
dilatational quadrant. 4 The compressional and dilatational
quadrants can be identified using seismograms recorded at
different azimuths around the source. The fault orientation and
a surface perpendicular to it can then be found, because in
these directions the first motion changes polarity. With the use
of additional data we can often tell which of these surfaces
was the actual fault. Given the fault orientation, the direction
of motion can also be found; note that the compressional and
dilatational quadrants would be interchanged if the fault had
moved in the opposite direction. The pulse radiated from the
earthquake also gives some information about the amount of
slip that occurred, the size of the area that slipped, and the
slip process.
Such observations of the location of earthquakes and the
fault motion that occurred in them are among the most important data we have for understanding plate tectonics, the primary process shaping our planet. The earthquake analyzed in
Fig. 1.1-7, for example, is like those that occur along the San
Andreas fault in northern California, part of the boundary
along which the Pacific plate moves northward with respect
to the North American plate. The fault is visible at the earth’s
Fig. 1.1-7 First motions of seismic P waves observed
at seismometers located in various directions about
the earthquake allow the fault orientation to be
determined.
“up”
Fault
DILATATION
COMPRESSION
DILATATION
COMPRESSION
Auxiliary plane
“down”
Epicenter
1.1 Introduction 5
science, words have multiple meanings.
5 In discussing analogous issues Sarewitz and Pielke (2000) note than even after billions of dollars spent on climate research, a senior scientist observes, “This may come
as a shock to many people who assume that we do know adequately what’s going
on with the climate, but we don’t,” and the National Academy of Sciences states that
deficiencies in our understanding “place serious limitations on the confidence” of
climate modeling results.
surface, so geological and geodetic observations also show the
motion that occurs in earthquakes. In less accessible areas
seismological observations provide most of the data used to
identify the boundary along which motion occurs and to demonstrate its nature. This is the case for most plate boundaries,
which occur in the oceans, beneath several kilometers of water.
Similarly, in subduction zones, where lithospheric plates
descend deep into the mantle and earthquakes can occur to
depths of 660 km, direct observations are not possible, but
analyses of seismograms reveal the motions and give insight
into their tectonic causes.
1.1.2 Models in seismology
As summarized in the previous section, seismology provides a
great deal of information about seismic sources, the structure
of the earth, and the relation of earthquakes to the tectonic processes that produce them. Even so, we will see that there are
major limitations on what the present seismological observations and other data tell us. For example, although we have
good models of seismic velocity in the earth, we know much
less about the composition of the earth and have only general
ideas about the deep physical processes, such as convection,
thought to be taking place. Similarly, although seismology provides a great deal of detail about the slip that occurs during
an earthquake, we still have only general ideas about how
earthquakes are related to tectonics, little understanding of the
actual faulting process, no ability to predict earthquakes on
time scales shorter than a hundred years, and only rudimentary
methods to estimate earthquake hazards. This situation is
typical of the earth sciences, 5 largely because of the complexity
of the processes being studied and the limits of our observations. Our best response seems to be to show humility in face of
the complexity of nature, recognize what we presently know
occurred is inferred from the three-dimensional pattern of radiated seismic waves. Figure 1.1-7 illustrates the method used for
an earthquake in which the material on one side of a vertically
dipping fault moves horizontally with respect to that on the
other side. This motion generates seismic waves that propagate
away in all directions. In some directions the ground first
moves away from the source (toward a seismic station),
whereas in other directions the ground first moves toward the
source (away from a receiver). The seismograms thus differ
between stations. In the “toward” (called compressional)
quadrants the first ground motion recorded is toward the receiver, whereas in the “away” (called dilatational) quadrants
the first ground motion is away from the receiver. Because the
seismic waves go down from the source, turn, and arrive at a
distant seismographic station from below, the first motion
is upward in a compressional quadrant and downward in a
dilatational quadrant. 4 The compressional and dilatational
quadrants can be identified using seismograms recorded at
different azimuths around the source. The fault orientation and
a surface perpendicular to it can then be found, because in
these directions the first motion changes polarity. With the use
of additional data we can often tell which of these surfaces
was the actual fault. Given the fault orientation, the direction
of motion can also be found; note that the compressional and
dilatational quadrants would be interchanged if the fault had
moved in the opposite direction. The pulse radiated from the
earthquake also gives some information about the amount of
slip that occurred, the size of the area that slipped, and the
slip process.
Such observations of the location of earthquakes and the
fault motion that occurred in them are among the most important data we have for understanding plate tectonics, the primary process shaping our planet. The earthquake analyzed in
Fig. 1.1-7, for example, is like those that occur along the San
Andreas fault in northern California, part of the boundary
along which the Pacific plate moves northward with respect
to the North American plate. The fault is visible at the earth’s
Fig. 1.1-7 First motions of seismic P waves observed
at seismometers located in various directions about
the earthquake allow the fault orientation to be
determined.
“up”
Fault
DILATATION
COMPRESSION
DILATATION
COMPRESSION
Auxiliary plane
“down”
Epicenter
1.1 Introduction 5