Beruflich Dokumente
Kultur Dokumente
7 M. BATH
SPECTRAL ANALYSIS IN GEOPHYSICS
8 O. KULHANEK
INTRODUCTION TO DIGITAL FILTERING IN GEOPHYSICS
9 T. RIKITAKE
EARTHQUAKE PREDICTION
16 P. MALISCHEWSKY
SURFACE WAVES AND DISCONTINUITIES
17 A.M.JESSOP
THERMAL GEOPHYSICS
Developments in Solid Earth Geophysics
18
ANATOMY OF
SEISMOGRAMS
OTA KULHANEK
Seismological Section, University of Uppsala, Uppsala, Sweden
ELSEVIER
Amsterdam — Oxford — New York — Tokyo
ELSEVIER SCIENCE B.V.
Sara Burgerhartstraat 25
P.O. Box 211, 1000 AE Amsterdam, The Netherlands
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any
form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the
prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions
Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands.
Special regulations for readers in the U.S.A.-This publication has been registered with the
Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive Danvers, MA 01923. Information
can be obtained from the CCC about conditions under which photocopies of parts of this
publication may be made in the U.S.A. All other copyright questions, including photocopying
outside of the U.S.A., should be referred to the publisher.
No responsibility is assumed by the publisher for any injury and/or damage to persons or property
as a matter of products liability, negligence or otherwise, or from any use or operation of any
methods, products, instructions or ideas contained in the material herein.
PREFACE
The primary goal of this book is to present in a rather tutorial form all the
necessary information and techniques pertinent to essential seismogram
interpretation. The treatment is descriptive rather than mathematical. Emphasis
is laid on practical aspects especially for the benefit of students and junior
seismogram interpreters affiliated with seismographic stations and
observatories. However, even workers more knowledgeable in seismology and
curious enough in the detailed deciphering of seismogram pecularities may
find the presentation useful.
with respect to possible travel paths through the Earth. The latter includes,
first of all, determination of basic source parameters (origin time, hypocenter
coordinates, size) but may also incorporate rather advanced studies (e.g.,
wave-form modeling, estimation of velocity distribution and moment tensor
determination). A large part of seismogram analysis is apparently a domain
of research and beyond the scope of this book. However, phase identification
is a doorway and obviously without correct seismogram interpretation hardly
any analysis would be possible.
The book is divided into two parts: a verbal description (Chapters 1-6) and
a collection of 55 plates (Chapter 7) with actual seismograms. The verbal
description explains in a rather elementary form the most fundamental
physical phenomena relevant to seismogram appearance. The collection of
plates exhibits a large variety of seismogram examples and corresponding
interpretations covering different seismic sources, wave types, epicentral
distances, focal depths and recording instruments.
The present book complements older manuals in that both analog and
digital records are considered. Seismograms from more traditional narrow-
band as well as from modern broad-band instruments are displayed. Tectonic
and volcanic earthquakes are represented and the exhibited seismograms form
a world-wide collection of records acquired from seismographic stations
located in North and Central America, Asia, Europe and New Zealand, i.e.,
in various geological and tectonic environments. Terminology and usage of
definition does vary among agencies in different parts of the world; that used
in this book is common in Europe.
This manual would not have been written without the support, continuous
interest and encouragement of Unesco. Invaluable indeed are the seismogram
VIII
Chapter 1
INTRODUCTION
In seismology (from the Greek word seismos meaning earthquake and logos
meaning science), irrespective of the final task, sooner or later the work
becomes dependent on seismogram interpretation, i.e. on discovery and
identification of recorded seismic waves. Guide-lines on seismic record
interpretation have long been requested especially by students and junior
analysts. My main objective here has been to serve these workers by
presenting a comprehensive and tutorial manual for deciphering available
seismograms, a work that has fascinated seismologists at all levels, from
genuine novices to legendary specialists, since the first seismogram appeared
in the early days of observational seismology, about one hundred years ago.
3
Chapter 2
Each year, planet Earth is shaken by some ten or more major and
destructive earthquakes killing thousands of people and imposing disastrous
economic consequences on the affected areas. If we reject the idea that
earthquakes are manifestations of God's displeasure, or caused by mysterious
dark forces, then what are the true causes of earthquakes and where do they
preferentially occur? Answers to these, and many other questions are
suggested by the current plate tectonics theory.
Plate boundaries are classified into one of the three following categories.
Boundaries where two plates are converging are called trenches; boundaries
where two plates diverge are called ridges; and boundaries where two plates
move horizontally past each other are called transform faults. Trenches are
also destructive plate margins. As two plates converge, one plate usually
bends beneath the other and descends into the soft, hot asthenosphere, a
process often referred to as subduction. The descending slab, also called
subduction zone or Wadati-Benioff zone, assimilates with the surrounding
mantle at a depth of about 700 km, i.e. approximately at the lower limit of
the asthenosphere, due to temperatures and stresses existing at that depth. The
locus of earthquakes in a subduction zone defines the Wadati-Benioff zone.
Ridges are also constructive plate margins. The opening where two plates
diverge is continuously filled by ascending mantle material.
4
Fig. 1. Present boundaries and relative motions of major tectonic plates. Arrows show the direction of
horizontal motions. Mid-oceanic ridges (heavy lines) are offset by transform faults (thin lines), hatchings
indicate subduction and dashed lines uncertain plate boundaries. (Based largely on the map of Press and
Siever, 1982).
From the description above, we would expect that the large majority of the
world's earthquakes are not generated randomly around the globe but in
relatively narrow belts along trenches, ridges and transform faults, i.e. along
boundaries between interacting mobile plates where deformation takes place.
A casual glance at a map of the Earth's division into rigid plates (Fig. 1)
and the global distribution of earthquakes (Fig. 2) confirms that this, indeed,
is the case. Some parts of the world are far more prone to earthquakes than
5
Fig. 2. Global geographical distribution of epicenters (circles) for earthquakes of surface-wave magnitude
equal to or greater than 6.7, during the period 1963-1988. The size of circles is proportional to the
magnitude of the events. Altogether, 1372 earthquakes are plotted. (Based on the map computer plotted by
W. Rinehart of NOAA).
According to their focal depth, earthquakes are classified into one of the
three categories: shallow, intermediate or deep. Shallow-focus earthquakes
(about 80% of the total activity) have their foci at a depth between 0 and 70
km and take place at oceanic ridges and transform faults as well as at
subduction zones. Intermediate-focus earthquakes (focal depth between 71
and 300 km) and deep focus-earthquakes (focal depth greater than 300 km)
occur at subduction zones. Most earthquakes originate within the crust. At
depths beneath the Moho, the number falls abruptly and dies away to zero at
a depth of about 700 km. Earthquakes along ridges usually occur at a depth
of about 10 km or less and are of moderate size. Transform faults generate
larger shocks at depths down to about 20 km. The largest earthquakes occur
along subduction zones.
7
SEISMIC SOVKCKS
Natural Events Man-Made Events
Controlled Events Induced, Triggered Events
Tectonic Explosions Reservoir Induced
Earthquakes Earthquakes
Volcanic Cultural Mining Tremors
Earthquakes Noise
Implosions, Collapse Fluid-Injection
Earthquakes Induced Earthquakes
Oceanic
Microseisms
Usually only the large and destructive shocks gain public attention. In fact,
the true number of earthquakes is much higher than most people realize.
Small events and earthquakes in remote unpopulated areas (e.g., under
oceans) usually occur without being noticed. The so called
magnitude-frequency distribution, studied in detail by many seismologists,
shows a rapid increase of the number of earthquakes with decreasing
magnitude. For the Earth as a whole, we have, on average, about 2 events
per year within the magnitude interval 8.0-8.9, about 15-20 events within the
interval 7.0-7.9, about 100-150 events within the interval 6.0-6.9, e t c The
International Seismological Centre (ISC), in Berkshire, England, lists about
25,000 to 30,000 events a year in its recent publications. Some workers
(B&th, 1979a) claim that there are as many as 1 million earthquakes in the
Earth every year, which means about two shocks every minute. Disastrous
earthquakes punctuate history. Annually, several tens of earthquakes cause
loss of life, property damage and environmental degradation.
9
Chapter 3
3.1 CRUST
3.2 MANTLE
Fig. 4. A segment of the Earth's interior, at two different scales, showing the location of major structural
discontinuities and the right proportions between crust, mantle and core. Numbers give the distance beneath
the Earth's surface in km.
3.3 CORE
Fig. 5. Distribution of compressional-wave (P) and shear-wave (5) velocities in the Earth's interior based
upon the Earth model CAL 6 computed at the University of California, Berkeley (Bolt, 1982). The low-
velocity layer at 100-250 km depth is clearly visible. S waves do not propagate through the liquid outer
core and consequently, the S curve is interrupted in the Earth's outer core. It is theoretically possible for S
waves to reappear in the solid inner core (by multiple conversion from Ρ to 5 and from S to Ρ at outer-
inner core boundary when entering and leaving the inner core, respectively), although such waves have not
been definitely observed.
13
Chapter 4
SEISMIC WAVES
When the strain accumulated in the rock exceeds its elastic limit a fault
ruptures, rock masses are abruptly displaced and seismic waves begin to
radiate from the fault. As the rupture propagates, it successively releases the
strain energy stored along the activated part of the fault. Thus, each point of
the fault contributes, with a certain time delay (due to the finite velocity of
the rupture propagation), to the total picture of seismic waves, which at a
certain distance from the causative fault interfere with each other and give
rise to quite a complicated wave train. At first sight, it seems that there is a
contradiction between the duration of the rupture at the source, which takes
between a fraction of a second and a few minutes, and the length of the
observed seismogram which for large and distant earthquakes can extend over
several hours. In fact, the length of the seismogram depends primarily on
various wave propagation effects such as reflection, refraction, conversion,
dispersion, etc and has very little to do with the duration of the quake.
Seismologists use the term coda to denote the part of the seismogram with
decreasing amplitudes which follows the principal phases.
Essentially, there are two types of seismic waves, body waves that
propagate through the Earth's interior and surface waves that propagate along
the Earth's free surface or along other discontinuities in the Earth's interior.
Surface waves carry the greatest amount of energy from shallow shocks and
are usually the primary cause of destruction that can result from earthquakes
affecting densely populated areas. Body waves radiated by the source
propagate in all directions (free waves) while surface waves start to
propagate first after body waves (different types) have been interacting along
boundaries. Thus, surface waves are always concentrated near discontinuity
surfaces and are, therefore, sometimes called bounded waves or guided
waves. In other words, for homogeneous media, i.e. for media with no
boundaries, there are no surface waves.
Body waves, which travel faster than surface waves, are of two types:
compressional (longitudinal) and shear (transverse). That an elastic body
should be able to transmit two different types of body waves was first
postulated by S.D. Poisson in 1829. At any given point of the body, the
velocity of propagation is determined by the density and elastic moduli at
that point. Compressional waves travel about 1.7 times faster than transverse
14
waves and are often called Ρ waves or primary waves (from Latin undae
primae). Transverse waves are frequently called S waves or secondary waves
(from Latin undae secundae). Ρ waves are always the first among seismic
waves that reach the recording station. Rock particles affected by a
propagating Ρ wave oscillate backward and forward in the same direction as
the wave propagates (see Fig. 6), analogous to, e.g., sound waves. In the case
of S waves, particles are displaced in planes perpendicular to the direction of
travel (Fig. 6) analogous to, e.g., light or electromagnetic waves. However,
since earthquakes generate Ρ and S waves, studies of recorded seismic waves
are, broadly speaking, more complicated than studies of sound or
electromagnetic waves. Fluids do not sustain shear strain, and therefore S
waves do not travel through liquid parts of the Earth's interior. Ρ waves, on
the other hand, propagate through both the solid and liquid divisions of the
Earth. Ρ and 5 waves had already been revealed on actual seismic records at
the end of the nineteenth century.
Fig. 6. Motion of rock particles (small arrows) that lie in the path of propagating P, S, LQ and LR waves.
Note that the waves propagate from the source to the receiver at the recording site, i.e. from left to right
in the sketch. Due to different propagation velocities, the waves will appear on the seismogram separated
in time in the order: P, S, LQ and LR. The large arrow indicates the direction of wave propagation.
discontinuities, through the crust and upper mantle. They can also travel by
different modes (overtones) which are often seen on records as
higher-frequency components superimposed on the surface-wave train. We
talk then about fundamental-mode and higher-mode surface waves. Higher
modes are most frequently observed for waves traversing purely continental
paths. In some cases, higher modes have also been associated with oceanic
paths. However, surface-wave higher modes disappear when the waves cross
the transition between continental and oceanic structures. Higher modes
propagate faster than the fundamental mode and are, therefore, recorded
ahead of LQ and LR.
etc. Generally speaking, as we depart from the focus, the wave amplitudes
diminish due to the anelastic attenuation (rocks are not perfectly elastic), by
geometrical spreading (the area of the wavefront increases with increasing
propagated distance) and by losses at interfaces (reflection, refraction, mode
conversion, diffraction, scatter). The anelastic attenuation is frequency
dependent (high frequencies are subject to high attenuation) so that
high-frequency seismic signals die out rather rapidly and may be recorded
only by proper instruments placed at relatively short epicentral distances. The
attenuation of geometrical spreading is frequency invariant.
Due to the internal structure of the Earth, at certain distance, e.g. around
20° or 144°, concentration (focusing) of energy of traveling seismic waves
takes place. Seismic signals recorded close to these distances often show an
increase of amplitude even with increasing distance, from the focus. This
phenomenon, which is limited to a few rather narrow distance intervals,
should be considered as an exception from the general behaviour of
amplitudes decaying with increasing epicentral distance.
Other factors influencing the amplitude of arriving seismic waves are the
source mechanism and the associated source radiation characteristics. Tectonic
earthquakes, in contrast to underground explosions, cannot be treated as
spherically symmetric point sources because the radiated seismic energy
transported by certain wave types is beamed in certain directions. Therefore,
two or more seismographs placed at the same epicentral distance but at
different azimuths with respect to an earthquake may, and most likely will,
show significantly different amplitudes of recorded seismic waves. It is also
likely that seismographs deployed at different azimuths will show different
amplitude ratios between arriving Ρ and S waves.
The plot in Fig. 7 serves to illustrate some of the basic properties of body
and surface waves described above. There is a sharp Ρ onset followed after
approximately three and a half minutes by a clear S onset (we neglect the
minor trace wiggles). About two minutes after the S arrival, we observe a
gradual increase of the amplitude due to the arriving LR wave
(vertical-component seismogram). At the beginning of the LR wave, the wave
period is about 40 s but it decreases to about 25 s after three or four swings,
clearly demonstrating the normal dispersive character of the recorded LR
wave. In this particular case (Fig. 7), the LR wave dominates the seismogram
indicating a shallow-focus earthquake.
Fig. 7. Seismogram of the earthquake in northern Greece on May 23, 1978, (magnitude M=5.7, focal depth
A=9 km), made at Uppsala, Sweden, at an epicentral distance of 2160 km. The trace has been made on a
long-period Press-Ewing seismograph (see Chapter 6) and shows the vertical motion of the ground. Time
advances from left to right and there is 1 minute between successive time marks (small upward offsets).
particular case, chosen for tutorial reasons, the noise level (cf the portion of
the record preceding the Ρ onset) is very low when compared with
amplitudes of recorded Ρ or S waves. Also, the decay of Ρ wave amplitudes,
so called Ρ coda, is rather rapid so that not only the Ρ onset but also the S
onset can easily be identified and the corresponding arrival times accurately
measured. Unfortunately, it is quite common that the analyst, in his daily
work, has to examine records with high background noise and to identify
various wave arrivals masked by noise, which often is a task in itself. This
may be especially true in cases of weaker and/or distant earthquakes.
Fig. 8. Examples of seismograms with oceanic microseisms. Upper part: strong microseisms recorded on
November 27, 1978 at Swedish seismographic stations Uppsala (left) and Umea (right). In both cases, the
traces are produced by standard long-period Press-Ewing seismographs and show the ground motion in the
E-W direction. While the noise level in the Uppsala record would still allow the seismic phases to be
picked, to identify weak arrivals in the UmeA record would certainly be a difficult task even for an
experienced interpreter. Lower part: seismic noise recorded on November 22-23, 1986 by a broad-band
vertical-component instrument at Grafenberg, Bayern, FRG. The analog monitoring (left) and the enhanced
analog display of the digital recording (right) are exhibited. The enhanced trace reveals the typical
dominant period of 6-8 s. (Grafenberg records provided by D. Seidl).
First, we shall consider waves from regional events, i.e. from events at
epicentral distances not larger than about 10°. For this range of distances, a
dominant portion of recorded seismic waves have propagated through the
crust and/or along the Moho discontinuity and are commonly called crustal
waves. The second category will include seismic waves recorded at an
epicentral distance between 10° and about 103°. Within this distance range,
seismograms are relatively simple and dominated by waves that have traveled
through the mantle. The travel paths through the crust in the vicinity of the
source (shallow focus) and the station are relatively short and often
considered of less significance for the total appearance of the seismogram.
The last category will treat seismograms obtained from epicentral distances
103° and larger. Records from these distances become complicated again and
contain waves (phases) that have traveled through the Earth's core (core
waves) or have been diffracted by the Earth's core. Earthquakes recorded at
distances less than 10° are called local events or regional events, while
shocks recorded from distances larger than 10° are called teleseismic events
or simply teleseisms. Some agencies refer to events between 10 and 20° as
regional and those beyond 20° as teleseismic.
Consider waves (rays) leaving the focus F and reaching the recording
stations 5 , S and S . Since the source radiates both Ρ and S waves, there
; 2 3
will be direct longitudinal and transverse waves recorded along the Earth's
surface. These waves have ray paths such as FS 2 (see Fig. 9) and are
encoded as Pg and Sg or sometimes as Ρ and S. The subscript g indicates
the travel path, which for seismic events in the upper crust (most of the
crustal earthquakes) is entirely confined to the granitic layer. A reflected ray
(e.g. ray path FRJSJ) is also possible from the Moho, and the corresponding
reflected Ρ and S waves are labeled as PmP and SmS, respectively. Note that
21
Fig. 9. Principles of wave propagation from the focus of earthquake F through a simplified one-layer
crustal model. Symbols Ο and Μ designate the Earth's free surface and Moho discontinuity, respectively.
S is the k-\h recording seismographic station, ι is angle of incidence, i, is angle of refraction, i is critical
k e
angle and V is velocity of propagation for Ρ or S. R are the points of reflection at the Moho discontinuity
k
for rays that travel to the k-ύι station. Ray paths are defined by points of origin, reflection and recording.
For example, FS is the ray between the focus and station S . For notation of crustal waves see the text.
2 3
in this case, it is only a part of the incident energy that is reflected from
Moho back into the crust and recorded at S . The rest of the energy is 7
refracted into the mantle and will never show up on the record made at Sj. It
follows from Fig. 9 that as the epicentral distance increases, the angle of
incidence i and the angle of refraction i also increase. At a certain critical
r
epicentral distance, i = 90°, which means that the energy of the refracted
r
ray does not penetrate into the mantle but travels along the Moho
discontinuity (cf the ray path FR R S ). The associated angle of incidence, /,
2 3 3
at distances shorter than the critical distance, which for the continental crust
is about 100 km.
Secondly, the true Moho and Conrad discontinuities are not planar and
strictly horizontal boundaries. In reality, they will show a certain dip and
some degree of undulation. Therefore, a two-layer crustal model with
somewhat irregular boundaries, as shown, in Fig. 10, will be more realistic
than that discussed above (Fig. 9). Further refinements of the model in terms
of additional discontinuities and their geometry will here be considered as a
domain of research rather than of an analyst in his daily seismogram
interpretation.
Accepting the model in Fig. 10, we realize that starting from a certain
epicentral distance, approximately 100 km, we record new phases, namely the
refracted Ρ and S traveling along the Conrad discontinuity. An asterisk in the
superscript position, P * and .S*, indicates this phase. An alternative code
sometimes used is Pb and Sb. The subscript b refers to basaltic layer.
At very short distances, less than 150 km or so, the first seismic wave
arriving at the recording station is Ρ or Pg, traveling with a velocity of about
6 km/s. For distances larger than critical but less than about 150 km, Pg is
followed by P * and Pn, in this order. P * and Pn travel with velocities of
about 6.6 and 8.0 km/s, respectively, i.e. significantly faster than Pg.
Therefore, at distances larger than approximately 150 to 200 km (depending
upon the true propagation velocities and thicknesses of the granitic and
basaltic layers) crustal waves change their order of arrival. For distances
larger than about 200 km, the first arriving phase is Pn, next arrives P * and
then Pg. Obviously, this is true only for continental travel paths.
Seismograms from earthquakes beneath the sea bottom, made at island or
coastal stations, will not show Pg or Sg phases since there is no granitic
23
Fig. 10. Principles of the wave propagation through a continental crust consisting of two layers with
dipping and non-planar interfaces. Symbol C designates the Conrad discontinuity and figures in the right
margin give approximate velocities of propagation in km/s for Ρ (upper figures) and S waves (lower
figures). Conventions as for Fig. 9.
Fig. 11. Principles of propagation of Pn, pPn and sPn waves. For the sake of graphical simplicity, an
one-layer crustal model is used. Conventions as for Fig. 9.
Fig. 12. Vertical (Ζ) and horizontal (Ν, E) component seismograms from a moderate size regional
earthquake. The event occurred off coast of southwestern Sweden on June 15, 1985 at focal depth of 15
km (magnitude M = 4.6). The traces are analog displays of broad-band digital recording (see Chapter 6)
L
made at Uppsala at an epicentral distance of 490 km. (Traces computer plotted by W.Y. Kim).
depth to the bottom. The depth region of low velocity in the ocean, called
SOFAR (sound fixing and ranging), provides extremely favourable conditions
for long-distance propagation of a special type of high-frequency seismic
wave.
Fig. 13. Short-period vertical-component record from a shallow (h = 33 km) earthquake in Norwegian Sea
made at Umea, northern Sweden. This earthquake occurred on November 21, 1967 (m = 5.4) at a distance
of 10° from Umea. The seismogram shows clear Ρ and S onsets, separated by 107 s. Approximately 6 min
after P, an onset labeled TSg is identified on the record. It corresponds to a wave propagating through the
water as a sound wave and, subsequent to a water-land conversion and refraction, as Sg over the land path.
In this particular case, the land path is about 1/3 of the total travel length. The TSg wave shows
oscillations with periods around 1 s and gradually increasing and decreasing amplitudes. The whole TSg
wave train lasts for about 2 min.
During the last 10 years, or so, observed Τ phases have proved very useful
in discriminating between underground nuclear explosions, detonated beneath
oceanic islands, and tectonic earthquakes (Adams, 1979). For this type of
explosion, the energy is injected directly into the SOFAR channel and
28
As follows from the above description, for epicentral distances less than
about 10°, the wave propagation is rather complicated. The seismogram
appearance varies from place to place due to regional variations in crustal
structure and consequently, for this distance range, it is difficult to list
generally valid clues for record interpretation. Nevertheless, some of the
following principles may guide the analyst to read correctly seismograms of
local and regional earthquakes.
1) Predominant periods of recorded crustal phases such as Pg, P*, Pn, Sg, S*,
Sn, etc are normally less than one second and hence best recorded by
short-period instruments. Rg periods are usually not longer than several
seconds.
2) It has often been observed that Sg has the largest amplitude (for cases
when large short-period Rg is missing), best seen on horizontal-component
records.
3) For epicentral distances less than about 200 km (depending upon the
crustal structure and focal depth), the first arriving phase is Pg. For larger
distances, Pn arrives first.
4) Near-surface events from distances less than about 600 km often generate
short-period Rg with clear dispersion, best seen on vertical channels.
5) Local and regional earthquakes of low or moderate magnitude are
characterized by short total record duration, usually not longer than several
minutes.
6) Island and coastal seismographic stations frequently record various kinds
of Τ phases.
Seismologically speaking, the mantle differs from the overlying crust also
in the fact that, in the first approximation, it may be considered as a laterally
homogeneous, i.e. as a spherically symmetric body. Seismic wave velocities
indeed increase with depth, however, the regional (lateral) irregularities,
typical for the crust, are almost absent (less distinct) in the mantle. Some
workers consider the distance range between 10 and 103° as ideal to record
not only the direct Ρ and S waves, but also the whole family of reflected
and converted waves. Travel paths of these waves are dominated by the
29
Body waves that lie entirely in the mantle and undergo no reflection
between the focus and the recording station are labeled with a simple
symbol Ρ or 5. Rays corresponding to travel paths of these direct waves (P
or 5), also called elementary waves or main waves, are displayed in Fig.
14. They depict paths of least travel time from the focus of the earthquake to
the recording site. Direct waves, when reflected one or more times from the
underside of the free surface, give rise to single or multiply reflected Ρ or S.
For example, the direct Ρ reflected from the free surface back into the
mantle once or twice, is called PP or PPP, respectively. In the same way,
we have also SS, SSS etc. Each letter, Ρ or S, in the symbol defines one leg
of the propagation path. Considering also the conversion from Ρ to 5, and
vice versa, on reflection, we may observe the wave denoted PS which travels
as Ρ from the focus to the reflection point at the free surface and from that
point to the recording station as S. PS and SP appear only at distances larger
than 40°. For a wave leaving the focus as Ρ and twice reflected/converted
from the free surface, we have four possible cases, namely PPP, PPS, PSP
and PSS. Some of these waves are sketched in Fig. 14. Obviously, we could
continue with three and more reflections/conversions and form the
corresponding wave symbols. However, from experience we know that it is
quite seldom that three and more reflections from the Earth's free surface are
clearly visible on actual seismograms. For distances larger than about 40°, the
30
Fig. 14. Examples of propagation paths of direct and reflected waves in the Earth's mantle. Solid and
dashed rays are used to distinguish between Ρ and S waves, respectively. Waves are generated by the
surface focus, F, of the earthquake which radiates both Ρ and S waves. Different shadings show the
mantle, outer core and inner core. For notation see the text.
Waves ascending from the focus to the free surface, where they are
reflected back into the mantle, are commonly called depth phases and are
denoted by a lower case prefix: ρ for longitudinal and s for transverse waves.
We can easily list the four possibilities of reflections near the epicenter,
31
which are pP, sP, pS and sS (Fig. 15). The first case, for example, denotes
the wave that traveled upward from the focus as Ρ (short leg) and had been
reflected back off the free surface again as Ρ (long leg). Depth phases,
primarily pP, are the most important phases routinely used in focal-depth
estimations. It is quite obvious that the deeper the focus, the later is the pP
phase in relation to P. Hence, accurately measured arrival-time differences
pP-P are reliable indicators of the depth of the focus. In the case of a deeper
focus, it is sometimes possible to recognize several different reflections from
the free surface. Such waves are then labeled pPP, pPS, pSP and pSS in the
case of waves with their short leg as P. Logically, sPP, sPS, sSP and sSS
denote corresponding waves with short leg as S (Fig. 15). Interpretation of
depth phases must be done with utmost care since, for example, pP from a
deep earthquake can easily be erroneously interpreted as Ρ when the first
arrival (P) is weak. Depending on focal orientation and other factors, sP may
be stronger than pP and may be mistaken for it. Depth phases are sometimes
stronger than the main Ρ wave, and may be the first readable phase.
Fig. 15. Examples of propagation paths of depth phases and their notation. Waves begin at the deep focus,
F, of the earthquake. Conventions as for Fig. 14.
The lower case symbol d (or its value in kilometers) inserted between PP,
SS, etc has been introduced by B.A. Bolt to indicate seismic waves reflected
from secondary discontinuities in the upper mantle. For example, symbols
P400P or P650P (Fig. 16) specify Ρ waves reflected at the underside of a
discontinuity at a depth of 400 or 650 km, respectively. These phases arrive
at the recording station ahead of the expected (calculated) arrival time for the
main PP phase and are frequently interpreted as reflections from upper
32
mantle discontinuities (i.e. as PdP). However, when the arrival time cannot
be explained in terms of known discontinuities as PdP, we call these onsets
early PP or precursors to PP.
Fig. 16. Examples of propagation paths of direct Ρ waves, Ρ waves diffracted around the core-mantle
boundary and Ρ waves reflected downwards at a discontinuity at 650 km depth. The discontinuity and the
shadow zone (103-144°) are shaded. Conventions as for Fig. 14.
reflected back into the mantle as PcP, PcS, ScP or ScS, while the latter,
called core phases, are refracted downward and enter the core. The refraction
is rather sharp because of the sudden significant drop of Ρ velocity beneath
the core-mantle boundary (see Fig. 5).
For initial ray paths which are only slightly steeper than the ray grazing the
core surface, corresponding PKP waves emerge at the Earth's surface at
distances beyond 180° (see Fig. 17). As the rays (surface-focus event) enter
34
the mantle more and more steeply, the core refractions become less and less
abrupt and the rays emerge at the Earth's surface at shorter and shorter
epicentral distances. This decreasing of distance stops at about 144°. Further
steepening of the initial ray paths results now in an increase of the distance
of emergence up to 165° or so. The phenomenon may be viewed in terms of
two PKP travel-time branches denoted PKP1 and PKP2 for the first and the
second arrival, respectively. Exactly at 144°, the waves from the two
branches coincide, the waves reinforce one another which gives rise to an
energy concentration near that distance. The point of largest energy
concentration is called a caustic point or simply caustic. It has to be
stressed that neither PKP1 nor PKP2 enter the inner core, i.e. both these
wave types have their deepest point of penetration in the outer core.
Fig. 17. Examples of propagation paths of Ρ waves traveling through the Earth's core. The shadow zone
between 103° and 144° is shaded and Β denotes the caustic point. Rays are numbered in the order of
increasing steepness of the initial descent. For details see the text. Conventions as for Fig. 14.
As we further steepen the initial ray path, we reach the family of rays that
enter the inner core (Fig. 17). These rays progress in a normal way, i.e. the
steeper the initial ray path, the greater the distance of emergence from about
110° until at last there is a ray that passes through the Earth's center and
reaches the Earth's surface at the antipode of the focus. Ρ waves that
traverse the inner core are denoted by /, giving rise to phases PKIPK,
PKIKS, SKIKS and SKIKP, although these are often still simply referred to as
35
PKP, PKS etc. Phases with an S leg in the inner core would include the
letter / , such as PKJKP, but these have never been unambiguously identified
on seismograms. For obvious reasons, both the symbols, / and / , have to be
accompanied on both sides by K. Rays corresponding to seismic waves
reflected at the outside and inside of the inner core are called PKiKP and
PKIIKP, respectively (Fig. 18).
Fig. 18. Propagation paths of Ρ waves traversing the Earth's interior from the focus F and reflecting at the
outside (PKiKP) or inside (PKIIKP) of the inner core. PKIKP is a Ρ wave refracted into the outer core
and through the inner core. Conventions as for Fig. 14.
In the region of the caustic, i.e. around 144°, the wave train of recorded
core phases becomes particularly complicated. It is first at distances beyond
the caustic point where observed onsets may be separated into individual
PKP branches. The energy distribution changes with the increasing distance.
PKP1 is the dominant branch just beyond the caustic, up to about 153°. In
records of weaker events (144-153°), PKP I is often the first visible onset
since PKIKP, theoretically preceding PKP I, is too weak to be observed. As
the distance increases, PKP1 becomes weaker and vanishes from records at
distances of about 160° and larger. For distances beyond, say, 157°, PKP2
usually dominates the seismogram. Some workers prefer the nomenclature
36
arrivals which are due to seismic waves reflected at the outside of the inner
core.
PKIKP in the distance range from about 125° to the caustic is often
preceded by early arrivals or precursors which can arrive many seconds
ahead of the main phase. These are best explained by scattering phenomena
at or near the core-mantle boundary.
In a similar way as above, we may form new symbols for the whole family
of waves propagating through the core. For example, PKKP is a Ρ wave
which has been reflected from the inside of the core-mantle boundary. PKKP
is often very pronounced on records made at distances between 60 and 80°.
The striking onset may easily be misinterpreted as a first Ρ arrival of another
event. Ρ waves trapped inside the Earth's liquid core and with multiple Κ
legs are called PmKP where τηΛ provides the number of reflections. Cases
like P4KP and P7KP have been reported (Bolt, 1982).
PKPPKP, or for short P'P', are PKP waves once reflected from the free
surface back to a station in the same hemisphere as the focus. Since PKP
has the caustic at 144°, one might also assume that the strongest reflection
will take place at that distance, and consequently the best chance to observe
P'P' is around distances of 2x144° = 288°, or 72° if we take the shortest
distance from source to station. P'P' is often well recorded, arriving about
30 minutes after the Ρ phase when most of the coda amplitudes of preceding
phases have already become faint, and it may in some cases be wrongly
interpreted as a new Ρ or PKP. 72° is also equivalent to 3 χ 144° = 432° or
(360 + 72)° so the phase P'P'P' is also strong at this distance, and may be
observed for strong earthquakes about another 20 minutes after P'P'.
Let us now shift our attention from Ρ to S waves. Similarly to PKP, there
are SKS waves, i.e. S waves traveling from the earthquake source down
37
through the mantle. Incident to the outer-core boundary they undergo a mode
conversion and as Ρ (the Κ leg) traverse the liquid outer core. Following
an inverse mode conversion, they again enter the mantle as S and emerge at
the Earth's surface as SKS. Analogous phases to PmKP are SmKS. First SKS
waves are observed at distances between 60 and 70° and the range of
observations extends out to distances of 180° or so. Depending upon details
in the structural model, SKS exhibits a caustic point at a distance of about
80° so that the best region to study SKS waves is that between 70 and 90°.
However, the phase identification has to be made with utmost care since SKS
waves recorded in this distance region are often contaminated with direct S
waves. At about 82°, SKS begins to arrive ahead of 5. For distances shorter
than about 95°, SKS is usually smaller than 5, however at distances beyond
95°, SKS amplitudes are often quite large. To mistake S for SKS and vice
versa will adversely affect the epicentral location. Since the epicentral
distance estimate is frequently governed by the observed arrival-time
difference (see Chapter 5) wrong 5 identification on the record will
provide a wrong epicentral distance which in its turn will result in erroneous
location. S and SKS are best recorded on long-period horizontal-component
seismograms. However, occasionally these body waves are also observed on
short-period records, although the onset time of the later of the two phases is
usually very emergent due to the contamination by the coda of the earlier
phase. An example of recorded S and SKS is shown in Fig. 19.
The period of SKS phases may reach several tens of seconds and is,
therefore, best recorded by long-period horizontal seismographs. On the other
hand, the best sensing of PKP, PKKP, PKIKP etc. is with short-period
vertical instruments.
Fig. 19. Analog displays of digital short-period (left) and long-period (right) seismograms from a deep-
focus earthquake in the Fiji Islands region recorded by the Chiang Mai station in Thailand, at an epicentral
distance of 90°. Both horizontal components are shown. The earthquake occurred on April 29, 1987, with
magnitude m = 5.9, at a depth of 390 km. At this epicentral distance the SKS precedes S. The short-period
records reveal a clear SKS, however the S arrival, about 30 s apart, is hidden in the SKS coda. The long-
period traces show both SKS and S and also the depth phase sS. After G. Choy (personal communication).
38
Fig. 20. Seismograms from a moderate size (magnitude m = 5.8), deep-focus earthquake recorded at
Swedish stations Umea (UME) and Uppsala (UPP) at epicentral distances of 60 and 64°, respectively. The
event occurred in the Sea of Okhotsk on February 1, 1984 at a focal depth of 580 km. The uppermost two
traces show the short-period Benioff (see Chapter 6) vertical-component record made at UPP. The middle
two traces exhibit the short-period S-13 (see Chapter 6) E-W component seismogram also made at UPP.
The bottom trace presents the long-period vertical-component record made at UME. Note the rather
impulsive character of recorded phases, in particular that of S and ScS, clearly visible in all three records.
Virtually no surface waves have been recorded from this event.
the Flores Sea area, on August 25, 1933, June 29, 1934 and June 30, 1943.
Their depths are given by Gutenberg and Richter (1938) as 720 km, although
other agencies have placed them rather shallower. ISC records contain four
recent events in the Fiji-Tonga area with depths greater than 750 km: January
15, 1981 (765 ± 17 km), November 21, 1982 (769 ± 31 km), October 25,
1972 (806 ± 84 km) and May 7, 1971 (848 ± 26 km). These events are all
small and not widely recorded, and their depths, particularly those of the
40
It can be shown that the amplitude decrease for body waves (P and S) is
inversely proportional to the propagated distance while for surface waves the
decrease in amplitude is inversely proportional only to the square root of the
distance traveled. Hence, with exception of very short epicentral distances,
surface waves carry by far the largest amount of wave energy radiated by
shallow and some intermediate-focus earthquakes. In Section 4.1, we showed
41
and 0.96 times the 5-wave velocity, v , of the half-space. For many rock
s
Fig. 21. Dispersion curves for group velocity for fundamental Love and Rayleigh waves that have traveled
along oceanic and continental paths (based on the diagram of Bullen and Bolt, 1985).
43
TABLE ι
Travel-time differences, T - T , and minimum periods,
Rmax F , for largest Rayleigh waves observed at
various epicentral distances Δ (from Β4th, 1947 and Willmore, 1979).
Δ Τ Δ Δ Τ
Λ
MM
1
MM
direct and anticenter routes are labeled Gl and G2, respectively. Waves
which have in addition traveled once around the Earth are denoted G3 and
G4, and so on. Accordingly, we have Rl, R2, R3, R4 etc. On many
occasions, observations of up to G8 and R8 have been made. As an
exceptional case we may mention records of the 1960 Chile earthquake,
M=8.3. Seismograms made at Uppsala, Sweden, reveal mantle waves G20
and R20 which have traveled a total distance equal to that from the Earth to
the Moon (Bath, 1979a).
47
Chapter 5
TRAVEL TIMES
One of the great seismological inventions during the first half of this
century was the establishment of highly accurate, relatively simple charts and
tables for times of travel of parent Ρ and S waves and of affluent families of
dependent body-wave types such as PP, PcP, PKP, SS, SKS etc. Travel-time
charts, also called holographs, and travel-time tables provide (i.e. predict)
within a few seconds the time required for a particular wave type to travel
from the hypocenter to a certain point on the Earth's surface, i.e. to a certain
seismographic station. Obviously, the travel time can also be viewed as the
time which passed between the instant of the wave generation and the arrival
of the wave at the station. The primary importance of travel-time tables for
observational seismology lies in the fact that they enable the analyst to
identify various arriving phases and to determine quickly the distance
between epicenter and station.
Body waves propagate through the Earth's interior rather than around the
surface of the Earth. Nevertheless, travel-time curves for teleseismic body
waves are always plotted as a function of arc distances expressed in degrees
(in this context, it is worth mentioning the Eiby-Muir tables, in which
distances are found for given time, rather than the other way round). As an
example, Fig. 22 displays travel-time curves of P, 5, Pc, PP and PcP waves.
As follows from the figure, travel-time curves for Ρ and S waves are
represented by two diverging curves. This means that the travel-time
difference T -T (hereafter S-P), increases with increasing epicentral distance.
S P
Fig. 22. Travel-time curves for P, Pc, PcP, PP and S. The travel time, 7, is given in minutes and the
distance, Δ, in degrees of arc from epicenter.
record onsets (phases) are indicated but they are also successively identified,
in agreement with their respective times of arrival by making use of available
travel-time tables. At the same time, it has to be emphasized that
identification of later phases based upon observed travel times alone may
lead to erroneous conclusions. For example, within certain distance ranges,
multiple rupturing generates later arrivals which may easily be mistaken for
pP or PcP. At about 39° distance, ScP (or PcS) and 5 from a surface-focus
earthquake arrive simultaneously and the correct identification becomes rather
puzzling. At distances around 82°, it is difficult to discriminate between 5
and SKS. A similar situation occurs for the individual PKP branches at
distances around 144°, and more examples can be found. In such cases,
reliable interpretation requires access to records from a number of stations
(located at different epicentral distances) and sometimes even a more
thorough study of the source physics. The latter evidently goes beyond the
usual duties of a station analyst. In practice, it is quite common that there
remain unidentified phases on seismograms in spite of the fact that many of
them were recorded with rather clear onsets.
Generally speaking, travel times of body waves depend upon both the
epicentral distance and the focal depth. By using the travel-time tables, it is
not difficult to demonstrate that for example, the travel-time differences like
S-P or PP-P are strongly distance dependent while the dependence on focal
depth is less distinct. On the other hand, for travel-time differences including
depth phases, such as e.g. pP-P or sP-P, the dependence is obviously the
opposite. Consequently, the former type of time differences will be used to
estimate the epicentral distance, whereas the latter will give the focal depth
of the recorded event.
It has been observed already in the beginning of this century when first
travel-time tables were constructed, that the time of travel for body waves
(for a given teleseismic distance and focal depth) are nearly the same
irrespective of the geographical region. Obviously, this would mean a
laterally homogeneous (spherically symmetric) structure of the Earth. Small
deviations, of the order of about 2 s, of observed travel times from those
listed in the tables are to a great extent due to structural deviations from
spherical symmetry (after allowance is made for the Earth's ellipticity). The
greatest systematic divergences have been observed between propagation
paths under the Pacific and under continental regions for distances of
approximately 50° and in continental regions between shield and mountainous
regions (Bullen and Bolt, 1985). Some workers advocate construction of
regional or even azimuth-dependent travel times (Bath, 1979a) to improve the
accuracy of teleseismic source locations. However, these and similar
proposals have as yet not materialized. Broadly speaking, in seismological
practice, we use global teleseismic travel-time tables considering a spherically
symmetric earth.
50
Since the turn of the century, when R.D. Oldham made the first separation
of longitudinal, transverse and surface waves in seismograms, a number of
travel-time tables has been constructed. Oldham produced provisional
travel-time tables for P, S and surface waves. The first widely used tables
were those of K. Zoppritz published in 1907. The subsequent evolution of
reliable tables during the first decades of this century has been a long and
laborious process including the work of H.H. Turner, B. Gutenberg, C.
Richter and others. In 1940, H. Jeffreys and K.E. Bullen published their
famous Seismological Tables (Jeffreys and Bullen, 1967), commonly
abbreviated as JB tables, which still serve as a standard for global events.
Jeffreys and Bullen started in the early 1930's from the so-called
ZdppritZ'Turner tables. After a series of successive approximations, the
resulting tables of 1940 provide times of travel accurate within the order of
one or two seconds for P, PKP and PKIKP and only a little less accurate for
a number of other phases (Bullen, 1975). The JB tables list times of travel
for Ρ and S as well as for all more common body waves. Curves
corresponding to the onset of Love and Rayleigh type surface waves are also
plotted. Times are given to the nearest tenth of a second, for epicentral
increments of 1° (0.2° for near earthquakes) and depth increments of 63 km
within the range from 33 to around 800 km. Times of travel for surface
earthquakes are listed as well (see Fig. 23).
TIMES OF Ρ - Continued
Depth h:
Surface o-oo OOI 0-02 003 004
m β m β 8 χη β m 8
30 6 οι·6 44*4
6 12-5 88 6 07-7 89 88 55-7 88 5 499 87 5 39-2
31 6 2ΐ·3 6 166 6 ιο·4 5 586 53·ι
88 88 87 θ4·5 87 86 5 47-8
32 6 Ι9-Ι οι·7
6 3θ·ι 87 6 254 87 87 Ι3·2 86 6 ο7·3 86 5 56-4
33 6 8·8 6 34· 1 6 27-8 2ΐ8 6 159 ιο·3
3 87 86 86 85 85 6 049
34 6 47.5 6 364 3θ·3 6 24-4 ι88
86 6 427 86 85 85 8 6 134
4
47-2
84
6 4ΐ·3 35-6 841 6
84 84 84
37 7 ΐ3·ο θ8·2 84 οι8
8 4
6 55-6 6 497 44·ο 6
84 84 84 83
38 7 2ΐ·4 ι66 ι ο ί 83 7 θ4·ο 6 58·ο 523 6
84 83 83 83 82 6
39 7 29-8 24-9 83 ι8·4 7 123 7 ο6·3 οο·5
83 83 82 82
40 7 38·ι 7 33-2 7 26-7 7 2θ·5 7 ΐ4·5 ο8·7 7 03.2
82 83 82 82 82
4i 7 46-3 7 4ΐ·5 7 349 7 28-7 7 227 ι6·ο, 7 ιι·3
82 82 82 82 8ι
42 7 54-5 7 49-7 7 43·ι 7 36-9 7 3θ·8 25·ο 7 ΐ9·4
82 82 82 8ι 8ο
43 8 θ2·7 7 57-9 7 5ΐ·3 7 45 ο 7 38-9 33·ο 7 274
8ι 8ι 8ι 8ο 8ο
44 8 ιο8 8 ο6·ο 7 59-4 7 53·ο 7 46-9 4ΐ·ο 7 35*4
8ι 8ο 8ο 8ο 79
45 8 ι8·9 8 ΐ4'θ
8ο 8 ο7·4 8 οι·ο 7 54-8 7 48-9 7 43-3
79 79 79 78
46 8 26 8 8 22-0 8 ΐ5·3 8 ο8·9 8 θ2·7 7 56-7 78 7 5ΐ·ι
79 78 79 78
47 8 34-7 8 2Q-8 8 23-2 8 ι6·7 8 ιο·4 8 <Η·5 77 7 58 8
79 79 78 78
48 8 42-6 8 37-7 8 3ΐ·ο 8 2 ·5 77 8 ι8·2 8 Ι2·2 8 ο6·5
77 77 77 4
6
8 ΐ9·8
7
Fig. 2 3 . Sample section (30° < Δ < 49°) of JB travel-time tables for Ρ waves. The left-hand column
indicates epicentral distance in degrees and the columns to the right give travel times successively for a
surface focus and for foci at 33, 96, 160, 223 and 350 km depth. Travel times are listed for depths (6370-
33) a + 33 where α is the fraction specified on the top of each column. First travel-time differences in 0.1
s/deg are given to the right in each depth column. (After Jeffreys and Bullen, 1967).
2 s (Bath, 1979a). Despite the fact that JB tables were constructed long
before the computer era, their impact on geophysics in the past 50 years or
so, can hardly be overemphasized. Large seismological centers like NEIC or
ISC still use the JB tables.
Herrin tables give travel times for Pg, Pn, Ρ, PP, PcP, PKP and travel-time
differences pP-P and PcP-P. Times of travel for P, Pg, and Pn are given to
the nearest hundredth of a second, the distance increment is 0.1° for Pg, Pn;
0.5° for Ρ and 1° for PP, PcP and PKP. Surface foci as well as foci at a
large number of various depths are considered (see Fig. 24).
For near seismic events, the construction of proper travel-time tables creates
a difficult problem due to crustal structural heterogeneities. Times of travel
52
TIMES OF Ρ — Continued
Depth i n kilometers
0 15 40 50 75 100 125
Μ S Μ s Μ S Μ s Μ S Μ S Μ S
Fig. 24. Sample section (37.5° < Δ < 45°) of Herrin travel-time tables for Ρ waves. Focal depth in
kilometers is given for each column, otherwise conventions as for Fig. 23. First travel-time differences in
s/deg are given on alternate lines in the Ρ table. (After Herrin et al., 1968).
for short distances (Δ < 10°) given in JB tables are based on observations
from Japanese earthquakes and times listed in Herrin tables are constructed to
be consistent with observed times in the central United States. Active
seismological observatories all around the world usually develop their own
travel-time tables which are appropriate to their respective regions. Two
examples of sample sections of travel-time tables for near seismic events are
depicted in Figs. 25 and 26.
53
Δ Pg Ρ* Pn Sg S* Sn Sg - P g PmP He
km m 8 m 8 m 8 m e m e m 8 m β m β m 8
0 0 0 0 0
10 1.6 2.8 1.2 3.3
20 3.2 5.6 2.4 6.6
30 4.8 8.4 3.6 9.9
40 6.4 11.2 4.8 13.7 13.2
50 8.0 14.0 6.0 14 .2 16.6
60 9.6 16.8 7.2 15.1 19.9
70 11.3 19 .6 8.3 16.1 23.2
80 12 .9 22 . 3 9-4 17 .2 26.5
90 14.5 25.1 10 .6 18.3 29.8
100 16.1 27.9 11.8 19.5 33.1
110 17 .7 18.7 20 .8 30.7 36.8 13.0 20.8 36.4
120 19.3 20.2 22 . 1 33.5 39.0 14 .2 22.1 39 .7
130 20.9 21.7 23.3 36.3 41.2 15.4 23.4 43.0
140 22.5 23.2 24.6 39.1 43.4 16.6 24.8 46.4
150 24.1 24.7 25 *9 41.9 45.6 17.8 26.1 49 .7
160 25.7 26.2 27 .2 44 .7 4 5 .9 47 .7 19.0 27.5 53.0
170 27.3 27 .7 28.5 47 .5 48.6 49.9 20.2 28.9 56.3
180 28.9 29.2 29.7 50.3 51.4 52.1 21.4 30.3 59.6
190 30.6 30.8 31.0 53.1 54.1 54.3 22.5 31.8 1 02 .9
200 32.2 32.3 32 . 3 55.9 56.8 56.5 23.7 33.2 1 06 .2
210 33.8 33.8 33 .6 58.7 59-5 58.7 24 .9 1 09.5
220 35.4 35.3 34.8 1 01.4 1 02 .2 1 00.9 26.0 1 12.8
230 37.0 36.8 36.1 1 04.2 1 04.9 1 03.1 27.2 1 16 .2
240 38.6 38.3 37 .4 1 07 .0 1 07 .6 1 05.3 28.4 1 19.5
Fig. 25. Sample section (0 < Δ < 250 km) of travel-time tables for near events with surface focus in
Sweden. Times of travel are listed for Pg, P*, Pn, Sg, S*, Sn, PmP, Rg and for the difference Sg-Pg.
(After Bath, 1979b).
54
NEAT? EARTHQUAKE P H A S E S
Times of Transmission for a Surface Focus
Δ Pg Ρ* Ρη Sg S* Sn
ο·ο ο ο·ο Ο (2-8) ο (6-8) ο ο·ο ο (3-9) ο (ιο·7)
0·2 4·ο 6-6
ο·4 8·ο (9-6) Ι3·2
ο·6 Ι20 ΐ3·ι ΐ9·8 2Ι·7
ο·8 ι6·ο ι6· 5 ι8· 3 20-4 27-7 3ΐ·ο
ι·ο 200 Ι9·9 2Ι·Ι 33·ο 33-6 361
Ι·2 239 23-3 239 39-6 395 4ΐ·2
ι·4 27-9 26-7 26-8 46-3 45-5 46-3
ι·6 3ΐ·9 30·2 29-6 52-9 5ΐ·4 5ΐ·3
ι·8 35-9 33-6 32-5 59-5 57-4 56-4
20 399 37·ο 354 ι 6·ι ι 3-3 ι ι·5
2-2 439 ο·4
4 38-3 Ι2·7 9-2 6-6
2-4 479 43-8 4ΐ·2 Ι9·3 Ι5·2 ιι·7
2-6 5ΐ·9 47-3 44·ο 259 2Ι·Ι ι6·7
2-8 559 5°·7 6·9
4 32-5 27· I 2Ι·8
3·ο 59-8 54· ι 497 39·ι 33·ο 20-9
32 ι 3-8 57-5 52-5 45-7 38-9 32·ο
3-4 7-8 ι ο·9 55-4 52-3 449 37·ι
3-6 11 8 44 58-2 58-9 5ο·8 42· I
3-8 ι ·8
5 7-8 I Ι·Ι 2 5-6 56-8 47-2
Fig. 26. Sample section (0 < Δ < 3.8°) of JB travel-time tables for near events with surface focus. Times
of travel are listed for Pg, P*, Pn, Sg, S* and Sn. Figures in parentheses give travel times of P*, S* and
Pn, Sn vertically reflected from Conrad and Mono discontinuities, respectively. (From Jeffreys and Bullen,
1967).
55
Chapter 6
Today, the market is flooded by a large variety of first class products and
key-ready instruments may be ordered for practically any type of
measurements the earth scientist may be interested in. To present an
exhaustive list and description of all past and existing types of seismographs
goes, therefore, far beyond the scope of this book and consequently we shall
limit ourselves only to a rather brief review of main principles and in
particular to those which have an impact on seismogram interpretation.
Most, but not all, of the seismometers employ the pendulum principle (see
Fig. 27). The pendulum's mass is connected through a vertical or horizontal
elastic suspension to a robust frame which in its turn is anchored in the
underlying bedrock. In more classical seismometers e.g. Benioff, Press-Ewing
56
electronic amplifiers. The only practical limitations are the seismic noise at
the recording site and the instrumental noise of the device employed.
Fig. 27. The principle of a vertical-pendulum moving-coil seismometer. Β is bedrock, Ρ is concrete pier, F
is frame, Μ is pendulum's mass, Ρ Μ is permanent magnet, C is coil, S is helical spring, Λ is amplifier and
FL is filter. Relative motion of the coil and the permanent magnet generates an electric signal that is
electronically amplified, filtered and recorded, together with time information, by a proper plotting device.
The arrows indicate that the instrument senses up-down, i.e. vertical, ground motion.
As the reader might have already realized, the terminology used above
distinguishes between instruments called seismographs and seismometers.
Older instruments such as Wiechert, Mainka, Milne-Shaw or Wood-Anderson
do not contain any seismic-to-electric transducers. Recording is made
mechanically or by mechanical-optical arrangements which are, for these
instruments, an integrated part of the whole apparatus. These instruments are
therefore called seismographs. On the other hand, more modern instruments
like Benioff, Grenet-Coulomb, Press-Ewing, etc. which work essentially as
seismic-to-electric sensors and can be coupled to various recording units, are
called seismometers. For the sake of completeness, let us add that devices,
nowadays mostly of historical importance, that indicate the occurrence of
ground vibrations but do not record their time dependence are called
seismoscopes. They may be considered as ancestors of seismographs with a
limited use for studying earthquakes.
59
and F again decreases. At very high frequencies of the ground motion, the
mass does not move at all which means that the relative movement (in
absolute value) between the mass and the frame is equal to that of the frame
and F = 1. It is evident, that in this case the recorded trace gives the true
picture of the ground motion. A simple sketch showing the dependence of F
upon the frequency of the ground vibrations is displayed in Fig. 28.
The relative period of the ground motion and of the pendulum is decisive
for the behaviour of the seismometer. We can distinguish three different
cases. First, if the pendulum's period is much longer than the period of the
ground motion (right-hand side of the diagram in Fig. 28), the recorded trace
is proportional to ground displacement. Second, if the two periods are about
the same (central part of the diagram in Fig. 28), then resonance will occur
and the trace is proportional to ground velocity. Third, if the period of the
pendulum is much less than that of the ground motion (left-hand side of the
diagram in Fig. 28), the trace is proportional to the acceleration of the
ground motion. Hence, by adjusting the pendulum's free period, we can
record various characteristics (displacement, velocity, acceleration) of the
ground motion.
Seismic signals may be very irregular and may occupy a broad spectrum of
frequencies and a wide range of amplitudes. Laboratory model measurements
use frequencies of 10 KHz and higher; industrial explosions recorded at short
distances produce signals with frequencies around 100 Hz; body waves from
regional earthquakes show frequencies in the range from one to several tens
of Hz; body waves from distant earthquakes usually have dominant periods
between 1 and 10 s and the same period range is also occupied by oceanic
microseisms; crustal and upper mantle surface waves have periods between
10 and 100 s whereas long period mantle surface waves often have periods
up to 10 s. The free vibrations of the Earth, generated by large earthquakes,
3
have periods of the order of one hour. "In no other branch of physics is there
such a great range in the frequency of the signals that are studied" (Bolt,
1982, p. 58). By measuring the actual seismograms, we may also estimate
the amplitudes of ground vibrations. For example, for regional earthquakes
the amplitudes may range from barely perceptible (small events) to several
centimeters (disastrous earthquakes). Amplitudes of surface waves from
61
Fig. 29. Period-dependent response characteristics for several seismograph systems: 1) Benioff (SP-
WWSSN); 2) Grenet-Coulomb; 3) Wood-Anderson; 4) Kirnos; 5) Wiechert; 6) Press-Ewing (LP-WWSSN);
7) Broad-band. SP and LP stand for short and long period, respectively.
62
TABLE 2
Basic parameters of some more common instruments
Benioff, SP (WWSSN) 1
1.0 0.7
Grenet-Coulomb 1.4 0.7 -10 4
Kirnos 22 80 -10 3
1
WWSSN stands for World Wide Standardized Seismograph Network
63
recorded waveforms. Displayed traces are vertical-component seismograms that would have been made at Erlangen, FRG, at an epicentral distance of
75° by (from top to bottom): 1) Short-period WWSSN seismograph; 2) Long-period WWSSN seismograph; 3) Kirnos seismograph; 4) Broad-band
seismograph. Relevant response characteristics are given in the left-hand margin of thefigure.After D. Seidl (personal communication), modified.
64
the four traces including various body-wave phases (deep-focus event), are
evident from the figure. The short-period narrow-band seismogram enhances
high-frequency (1-3 Hz) wave trains, an advantage employed for accurate
determination of onset times of various arriving phases. On the long-period
WWSSN, Kirnos and broad-band displacement seismograms, the
high-frequencies have been filtered out and low-frequency waveforms
dominate the records which are suitable for 5-wave studies, determination of
dynamic source parameters, construction of synthetic seismograms, etc. While
the high-frequency signals reflect details of the earthquake source process and
the fine structure along the propagation path, the low-frequency components
reveal the large scale source and structural characteristics.
Bayern, FRG. The top two traces are computer simulations of corresponding
short-period and long-period seismograms that would have been produced by
WWSSN systems placed at Grafenberg. The earthquake occurred at a
distance of 11° and at a depth of 112 km. It follows from the figure that the
standard short-period instrument separates various recorded wave groups,
whereas the long-period system accentuates the general record features at the
cost of smoothing (filtering) out the fine structure of the record. The 1977
Romania quake was a double shock which is well demonstrated in the short-
period and broad-band traces where a double Ρ phase, labeled PI and P2,
with a separation of about 5 s, is clearly visible. Observe that this feature is
hardly discernible in the long-period record. The strong pulse arriving
approximately 10 s after PI is the so called stopping phase attributed to the
termination of the rupture process. It is characterized by a polarity opposite
to that of the initial PI pulse.
converter which transforms the signal into a corresponding digital format. The
advantage of digital instrumentation and recording media is the ease with
which the stored data can be further processed by means of digital
computers. We also gain from significant enlargements of the recording
dynamic range, i.e. from the ability of digital equipments to record with one
system both very large- and very small-amplitude signals.
Fig. 32. The Mexico earthquake (magnitude Μ = 8.1) of September 19, 1985 (upper trace) and the largest
aftershock (magnitude Μ = 7.5) of September 21 (lower trace) recorded at Matsushiro, Japan, at an
epicentral distance of 101°. Long-period, vertical-component analog displays of digital records are exhibited.
Enlarged intitial segments, the first 6 minutes or so, of the records from both events, are also shown. For
more details see the text. After M. Yamamoto (personal communication), modified.
the data volume may be significantly reduced by saving only selected seismic
signals and deleting the rest of the record. At present, many observatories
equipped with necessary computer facilities and magnetic-tape recording use
this approach of archiving data and produce so called event tapes. In spite of
continuous technical improvement, to store complete digital records of
short-period data is still prohibitively costly. Also, the traditional
photographic paper recording which does not allow a similar data reduction
becomes more and more expensive requiring at the same time larger and
larger archiving facilities.
68
Fig. 33. Seismogram traces from a microearthquake in northern Sweden on October 19, 1988, recorded at
an epicentral distance of 80 km. Recording was performed by a mobile temporary seismographic station
equipped with a short-period seismometer and a digital recording system. Exhibited traces show E-W
component analog displays of digital records. From top to bottom there are: 1) original trace containing
frequencies up to 80 Hz; 2) original trace filtered by means of a low-pass filter (sixth-order Butterworth)
with 12 Hz cut-off frequency; 3) segment of the filtered trace with extended time scale. (Traces computer
plotted by S.O. Under).
Digital records stored on proper recording media are fed directly into the
memory of a digital computer. The computer is programed to perform much
of the routine work which earlier was the sole domain of the interpreter.
Nevertheless, at well established observatories it is quite common to operate
analog, often visual, monitoring systems in parallel with the digital recording.
It seems that individual scrutiny of the analog seismograms by an
experienced interpreter is still essential for efficient use of digital recordings.
69
Chapter 7
SEISMOGRAMS AND INTERPRETATIONS
TABLE 3
Nomenclature of seismic phases
Symbol Meaning
Teleseismic events
TABLE 3 (continued)
Symbol Meaning
PKIKP, (or P", PKP ) DF Ρ wave traversing the outer and inner core.
PKiKP Ρ wave reflected at the boundary of the inner core.
PKIIKP Ρ wave reflected from the inside of the inner-core boundary.
PKKP Ρ wave reflected from the inside of the core-mantle boundary.
PmKP (m=3,4,...) Ρ wave reflected m-1 times from the inside of the core-mantle boundary.
SmKS (m=3.4,...) S wave converted into Ρ on refraction at the outer core, reflected m-1
times from the inside of the core-mantle boundray and finally converted
back into S when again entering the mantle.
PKPPKP (or P'P') PKP wave reflected from the free surface, passing twice through the core.
P'dP' PKP reflected at the underside of the discontinuity at depth d in the upper
part of the Earth, d is given in kilomters.
LR Surface wave of Rayleigh type.
73
Μ = 7.2
TABLE 4 (continued)
74
Date Origin Geographical Epicentral Focal Magnitude 1
Comments 2
75
76
TABLE 4 (continued)
1
m, Μ and M stand for body-wave, surface-wave and regional magnitude, respectively
L
2
p.c. = personal communication
77
TABLE 5
Seismographic stations used
TABLE 5 (continued)
Ν
Is' A^/yy'',AA'< J W\Vw«VVV , , V' J "VT]
Λ II
Pit Ρ2,
A •--Λ^υν^Λ''Μ vvvwv· .Λ/\,-.
Slj S2 Ε
ΜΛ ifV'i/Vv-
Plate 2. Direct P and 5 waves from a double microearthquake in western Bohemia, Czechoslovakia,
recorded by a temporary short-period digital seismographic system deployed at an epicentral distance of 5
km. The two events occurred on January 21, 1986 at a focal depth_of 8 km and were separated in time by
less than 1 s. The events have ML magnitudes of 1.4 and 1.5. PI and P2, which are best seen on the
vertical channel, denote P waves from the first and the second shock, respectively. Similar notation is used
for S waves which can easily be identified on both horizontal channels. There is 1 s between successive
time marks at the bottom of the plate. According to the ISC solution (Table 4), the epicenter was in
Germany and the focal depth was assumed to be 10 km.
00
00
Plate 3. Example of an earthquake swarm recorded at Kamata (KTJ) by short-period seismographs at an epicentral distance of about 10 km. The
swarm occurred at shallow depth near Izu peninsula, Japan, on October 14-15, 1984. The attached map shows the location of KTJ with respect to the
epicentral area (shaded). The displaced traces are, from top to bottom, N, E and Z components. The two horizontal components exhibit displacement
whereas the vertical component presents velocity of ground motion. During the 5-6 hours of records displayed in the plate, several hundreds of events
were recorded. Observe that due to the short epicentral distance, the S-P arrival-time difference is of the order of 1 s. Thus, in many cases, it is not
possible to separate on the record P arrivals from corresponding S arrivals.
6.00 7.00 8.00 9.00 10.00
Plate 4. Regional earthquakes of October 9, 1986 at Sierre and of July 7, 1985 at Langenthal, Switzerland,
demonstrating the main features of seismograms obtained at a regional short-period station network.
Locations of the recording stations and epicenters are depicted in the attached map. Note that only stations
LLS, MUO, SLE and TMA belong to the standard Swiss network. Number 5 in the map denotes the
earthquake near Sierre which occurred at a depth of 5 km while number 32 indicates the event near
Langenthal which has a focal depth of 30 km. Displayed sets of seismograms show the development of
basic crustal waves, like Pg, (P), Pn, PmP and corresponding S waves, for epicentral distances between
approximately 5 and 180 km. Records to the left belong to the Sierre event and records to the right to the
Langenthal event. The records are ranged in such a way that they can directly be compared with travel-
time curves (central part of the plate) and corresponding ray-path graphs (bottom part of the plate). The
time axes in all graphs are reduced by the factor distance/6 to enable a compact display. The displayed
seismograms represent the vertical movement of the ground. Maximum amplitudes are normalized for better
visibility in the recordings and therefore do not reflect the true ground motion at the stations. Real
amplitudes vary according to the actual magnification of recording systems.
In each seismogram the identifiable phases are indicated and different branches in the travel-time curves
are labeled accordingly. The most prominent phases for distances larger than 80-100 km are wide-angle
reflections at discontinuities in the lower crust (PmP, SmS).
Note, that the velocity-depth functions used to calculate the tavel-time and to plot the ray-paths are varying
laterally. This is typical for the heterogeneous crustal structure beneath the network in Switzerland and in
many other areas. The crustal thickness in Switzerland varies from 26 km (préalpine region) to almost 50
km (alpine region) for the example shown here. For the calculation of travel times between the source and
stations, this variable crustal thickness has to be considered. Flat uniformly layered velocity models would
give large uncertainties in the determination of the source parameters.
-™^^*ΛΛ^^^ ÎM^f**^^
Plate 1. Direct P and S waves from a local microearthquake, AfL=1.8, recorded at Matsushiro, Japan, on
July 11, 1986. The epicentral distance is 5.4 km and the focal depth 6.9 km. All three components
recorded by a short-period seismograph are displayed. Due to the short epicentral distance, the waves travel
close to vertical (see the inset) which in its turn results in large P amplitudes on the vertical component
and significantly smaller amplitudes on both horizontal components. Obviously, for S, the relation is the
reverse. The short epicentral distance does not permit the other crustal phases to develop. There are 2 s
between successive time marks at the bottom of the plate.
83-84
iOO
DISTANCE (km)
25
(_)
LU
tn 2 0
z
~15
LU
-10
LJ Pg,PmP
E o
LJ
tr
100 200
DISTANCE IN KM
100 200
85-86
Ι**+^****~+^*^**^**-*+**·
Plate 7. In 1978, the Japan Meteorological Agency, JMA, established a permanent ocean bottom
seismograph observation system, TKOBS, off the south coast of the Tokai District in central Japan. Four
stations were set up on the continental shelf at water-depth range between 0.7 and 2.2 km. Each station is
equipped with three-component short-period seismometers. The orientation of the horizontal instruments
depends upon the direction of the seismograph vessel at the time of installation. The plate exhibits short-
period three-component earthquake records made at the TK02 station, which belongs to TKOBS, at an
epicentral distance of 111 km. The earthquake took place off the south coast of the Tokai District, central
Japan, on January 10, 1981 (A/ =3.3) at a focal depth of 40 km. The vertical component (bottom trace)
L
shows clear Ρ and Τ wave trains, while the two horizontal components (upper two traces) display clear S
and Τ phases. The Τ wave is an acoustic water wave traveling through the SOFAR channel in the sea with
a velocity of about 1.49 km/s. Time separation betwen successive vertical lines is 4 s.
88
" 1 •U r ~ Γι 11 α λ - ^ -
-υ-η Π
In spite of comparable magnitude and hypocentral distances, the two records look differently. As is
generally the case, Ρ and S phases are more pulse like and better defined for the deeper event B. The
seismogram of this event also shows less complexity between the Ρ and S phases when compared with that
from the shallow earthquake A.
89
1 1 1
Plate 9. Short-period seismograms from a deep-focus earthquake on November 16, 1985, west of Noto
peninsula, Japan, made at the Matsushiro station at an epicentral distance of 157 km. Since the focal depth
of this event is 308 km, the direct waves are incoming almost vertically (see the inset). For an event of
this focal depth, one expects rather simple impulsive Ρ and S waves on the seismograms. The Ρ arrival is
quite distinct on all three components while the S onset is best seen on the E-W component. There is 20 s
between successive time marks at the bottom of the plate.
90
Plate 10. Velocity (left) and displacement seismograms (right) of the Swabian Jura, southern FRG, earthquake of September 3, 1978 (M =5.9). A three-
L
component broad-band recording system at GRF, Bayern, FRG, at an epicentral distance of 220 km, has been used. All four essential crustal phases,
i.e. Pn, Pg, Sn and Sg can be identified on the records, even though the Sn onset is somewhat ambiguous. Sg is best seen on the horizontal channels.
The enlarged segment of the initial portion of the vertical-component displacement record (framed) shows a clear "one-sided" onset between Pn and
Pg, arriving 2.6 s after Pn. By making use of synthetic (computed) seismograms, this phase has been interpreted (Kind, 1979) as sPn (see Fig. 11).
The successful revelation of the sPn phase made it possible to determine the focal depth of the event to be about 6 km. As follows from the displayed
records the velocity seismograms enhance the high frequencies and thus are useful in precise readings of the onset (arrival) times. The displacement
seismograms, on the other hand, emphasize the low frequencies and are suitable for determination of dynamic source paramters, for displaying the low-
frequency surface waves, etc. There is 10 s between successive time marks at the bottom of the plate.
91
Horizontal 1
Horizontal 2
PR 2
SEA SURFACE
CRUST
Plate 11. The displayed seismograms were made at TK02 which belongs to the TKOBS system (for details
see Plate 7). The onset of Ρ is clearly seen on the vertical channel together with PR1 and PR2 which are
waves reflected from the sea surface (see the sketch). This earthquake occurred on July 27, 1987 at an
epicentral distance of 247 km and at a focal depth of 156 km. The sharp, impulse like, Ρ on the vertical
component is a typical record characteristic of a deep- or intermediate-focus shock. The S arrival is best
seen on the horizontal channels.
92
Plate 12. The disastrous Friuli, northern Italy, earthquake of September 15, 1976 (m=5.4), recorded by a three-component broad-band system at GRF,
Bayern, FRG, at an epicentral distance of 400 km. Segments of velocity seismograms (left) and complete displacement seismograms (right) are
displayed. Three clear onsets corresponding to Pn, Pg, (vertical-component, velocity record) and Sg (horizontal-components, displacement records) are
visible on the exhibited records. Observe the dominant, fundamental mode, LR wave with a distinct dispersion, in the period interval from about 15 to
5 s, best developed on the vertical-component channel. The short-period oscillations preceding LR belong to the higher mode Rayleigh waves. The
dominant surface wave is a strong indication of a shallow event. Researchers still disagree on the more precise focal depth of this shock. Reported
estimates vary from 5 km to more deeper parts of the crust. There is 10 s between successive time marks at the bottom of the plate (note, there are
different time scales for the velocity and displacement traces).
93
Plate 13. Short-period three-component records from two earthquakes (A and B) east of Honshu, Japan,
demonstrating the influence of the source mechanism. Source parameters of the two shocks which both
took place in the subduction zone at a depth of 35 km are the following:
The recording site is in both cases the Matsushiro station, Japan. The first onset, probably Pn, is clearly
seen on records from both earthquakes, but otherwise the two cases are quite different. Observe that the
top three traces (event A) are dominated by relatively low frequencies while the three bottom traces (event
B) consist of waves with rather high frequencies. Due to the proximity of the two hypocenters the effect of
different travel paths can be disregarded and the striking variation in the frequency content is then ascribed
to different source mechanisms. Arrivals of Sn and Sg are expected 45-55 s behind Pn. There are 2
minutes between successive time marks.
94
Plate 14. The upper trace shows a record of clear Ρ and S and well developed Τ phase from an earthquake in the Tonga Islands region on August 20,
1986 (m=5.5, A=86 km). Exhibited seismograms were made at Niue, at an epicentral distance of 463 km, by a short-period, vertical-component
seismograph. Niue is an atoll some 20 km in diameter whichrisessteeply from surrounding ocean floor some 5 km deep. The relatively long length
of the Τ wave train is due to the intermediate focus of the earthquake. In this particular case, the energy has entered SOFAR over a considerable
length of the path between the epicentre and Niue (compare with the short duration of the Τ wave produced by a shallow underground explosion,
displayed in the lower trace). The lower trace shows a seismogram of a Γ phase from an underground nuclear explosion in the Mururoa Atoll in
French Polynesia recorded at the island station Rarotonga at an epicentral distance of 19°. The explosion was detonated on October 26, 1985 (m=5.4)
and the instrument used was a short-period vertical component Benioff (WWSSN). Τ waves are very useful in identifying the underground nuclear
explosions because their amplitudes are up to 30 times greater than those of the associated Ρ waves. Characteristics which distinguish Τ waves from
nuclear explosions recorded at Rarotonga include: 1) A very sharp onset of maximum energy, a characteristic of shallow sources. The shallow depth of
the explosions (presumably about 1 km) and the fact that they are denoted beneath oceanic islands, would allow energy to be injected directly into the
low-velocity sound channel (SOFAR) in the ocean. 2) There are characteristic later arrivals in the Τ wave train (marked R on the seismogram) which
are interpreted as 7-wave reflections from islands near the path from Mururoa to Rarotonga (for more details see Adams, 1979).
95
Plate 15, Three examples of short-period recordings from Α-type volcanic earthquakes. The top and middle
traces display records of quakes from the Asama volcano, Japan, on June 6 and 30, 1985, respectively
made at the Asama Volcanic Observatory at a distance of 3.2 km from the crater. E-W components are
exhibited. The bottom trace shows the vertical-component seismogram of a volcanic earthquake from the
Tacan£ volcano, Guatemala, on May 30, 1988. This event was recorded by the permanent station Sibinal at
a distance of about 6 km from the crater. The recorded traces show high-frequency oscillations, sharp
onsets and a gradual amplitude decrease forming a typical delta-shape envelope. Ρ and S waves arrive
within one second which restricts the focal depth to several kilometers.
96
* ''ΐ|Π| ·
Μΐ 1
»ι,:Ι'ψ'|'Ι |,
- " . · ' Ή f™ ι I - · I·
Plate 16. Short-period seismogram examples of Β-type earthquakes. Top traces (vertical component) show a
number of quakes from the Poas volcano, Costa Rica, recorded on April 27, 1988 at Poas station located
3.5 km from the crater. The activity is due to the degasification of magma. Bottom traces are recordings
(E-W component) of earthquakes from the Asama volcano, Japan, obtained on March 21, 1983 at the
Asama Volcanic Observatory at an epicentral distance of 2 km. When compared with the Α-type volcanic
earthquakes (see Plate 15), the above seismograms are characterized by longer periods and by the absence
of distinct first onsets. The duration of oscillations is usually less than 30 s. It is assumed that these
earthquakes originate at depths lesser than those of the Α-type shocks.
97
Plate 19. Three examples of volcanic tremors recorded by short-period instruments in the immediate
vicinity of active volcanoes. The top and middle traces exhibit the vertical-component while the bottom
traces display the E-W component of the ground motion. Volcanic tremors are most likely generated by
lasting oscillations of the source volume due to injection of new material, by movements of liquid magma
and by boiling effects in the magma chamber or in a conduit The top traces are harmonic low-frequency
tremors from the Rincon de la Vieja volcano, Costa Rica, recorded at station RIN on May 15, 1986 at a
distance of 6.4 km from the crater. Tremors from the Arenal Volcano, Costa Rica, recorded on March 27,
1987 at a distance of 3 km from the crater can be seen in the middle traces. The bottom traces show
volcanic tremors from the Asama volcano, Japan, observed on February 5, 1983 at the Asama Volcanic
Observatory at a distance of 2 km from the crater. As follows from the displayed traces, volcanic tremors
may appear on seismograms as more or less monochromatic signals (top and bottom traces) but also as
signals with varying, i.e. low and high, frequency (middle traces). The signal duration is rather long,
lasting for up to a day or so.
100
Plate 20. The top trace shows a classic example of a sonic boom recorded at Priest Mountain, California,
on April 14, 1981 on a short-period vertical-component seismograph. The sonic boom was caused by the
re-entry of the space orbiter Columbia at Edwards Air Force Base in southern California. The sonic boom
was visible on nearly all of the central and northern California stations in the Berkeley network. Sonic
booms travel at the sound velocity in air which is 285 m/s at sea level. The typical signature of a sonic
boom recorded on short-period instruments is a gradual buildup of record amplitude culminating with a few
swings at very large amplitudes when the shockfront passes over the station followed by a gradual decay
of motion. The duration of the signal is typically 30-150 s.
Bottom traces display a pair of similar unknown-source events which were recorded on the short-period
vertical-component at Mina, Nevada, on October 22, 1986. These are typical examples of similar events
which are occasionally recorded at Mina. It is possible that these records are generated by supersonic
aircraft flying at low altitude although this has not been confirmed. The signature is roughly similar to the
signature of some sonic booms caused by supersonic flights observed on seismograms made in the central
coast of California (see e.g. the top traces).
101
Plate 2 1 . Two examples of ripple charge blasting at a local quarry recorded on a short-period vertical
component seismograph at Jamestown, California (top traces). The quarry is 16 km north from Jamestown
as determined from the Rg-Pg onset time difference of 2 s. The equivalent local magnitude for the larger
of the two blasts is M = 1.5. The short-period Rayleigh waves typically start at about 2 Hz and exhibit
L
dispersion. The dominating Rg phase indicates a near-surface event. One distinguishing feature between
small quarry blasts and small natural earthquakes recorded at short distances in the Sierra Nevada in
California on high-gain instruments is the dominant period of the largest amplitude phase which is
generally around 2 Hz for blasts and higher than 5 Hz for natural earthquakes.
The bottom trace exhibits an example of a cavity collapse following an underground nuclear explosion at
the Nevada Test Site, recorded at Priest Mountain, California on November 14, 1986 at a distance of 480
km. The collapse event occurs when the chimney above the cavity created by the explosion begins to fall
in the cavity. The collapse typically takes several seconds to a few minutes and it very efficiently excites
omnidirectional Rayleigh waves because it is like a point downward force on the free surface. This
example is typical of a collapse signature in that no clear onset is evident and the duration is of the order
of 5 minutes (not seen in the displayed record) with the largest amplitudes near the middle of the record.
103—104
Plate 22. Long-period three-component seismograms of the earthquake on August 15, 1985 in northwestern
Hungary (Af=5.0, A=10 km) made at Uppsala, Sweden, at a distance of 13°. Displayed records show the 5"
arrival, best seen on the E-W component, and a remarkable continental-path surface-wave dispersion. The
waves arrive at Uppsala almost exactly from the south and hence, the large amplitudes on the E-W
component were interpreted in terms of LQ waves arriving ahead of LR. P waves are in this case not
discernible on long-period records.
-WV^v^^
t»y>^^%»^»^myi
Plate 23. Long-period records of the September 19, 1979 earthquake in Umbria, Italy (Λ/=5.9, h=\6 km),
made at Toledo, Spain, at an epicentral distance of 13.2°. Vertical and E-W component records are
displayed. Distinct P and S onsets are seen on the seismograms. The large amplitudes of the LR wave train
exhibit very clear dispersion. Higher modes are not visible, most likely due to a mixed continental and
oceanic travel path and to complicated crustal structure.
u
W
L·øm*m^^,*w**4~+*»"
* 0 * ^ * *ã******Μ« WwMM M WH*m#*#H *** ****
» * f » # S
» I · * v
"'{Wf^"/\^\
Plate 24. Interesting records from a near-surface earthquake in the Iceland region (m=5,3, A=l km), made at Uppsala, Sweden, at an epicentral distance
of 19°. The earthquake occurred on September 15, 1973. Long-period three-component seismograms are displayed. The vertical component exhibits a
weak P onset whereas the S arrival is very distinct in all three channels. There are well developed trains of Love and Rayleigh waves which can be
seen in the horizontal and vertical components, respectively. Observe that LQ starts right at the arrival of 5. All three channels show a clear dispersion.
In this particular case, the waves traveled along a mixed, oceanic and continental propagation path.
w*w~y**y^^
x
" * * ' ^
~ ^ ^
Plate 25. Broad-band three-component, displacement seismogram from the Norwegian Sea earthquake on July 27, 1976 (m=5 1 A=9 km) The records
were made at GRF, Bayern, FRG, at an epicentral distance of 23°. All three channels show simple and clear patterns of P and S waves' with distinct
onsets as well as of dispersed surface waves. Note that whUe the vertical channel displays the Rayleigh wave train, Love waves which precede the
recorded LR by one minute or so, are well recorded on the E-W component. The strong surface waves indicate a shallow focus event There is 1
minute between successive time marks at the bottom of the plate.
o
^1
o
109—110
J?m
ÉƺΡ: _
m
- " " — /^ A _ â
-Z^-
~Ρ :sz
tfJfaAjüfl·*^^ <Á**%·^ΛÌ***^í «^ÌÌ^^Λ^í* ty*Aj\^
J?-
HIP
4\jlmAm4m* LR: ^e.
t »»tf^ 23 -I
*A**\^'
^ΛÀ^
J^
\j\/\f\J\l\ft
w ™ ™ ^ v w»
W y » i » i IMÎ'I '
A^w*^N^W»w"«»w .^AA^^AAMM^^W
Zll5*0.
I » ^ * ^· ^ * ·
-L
íË/\ÃÍ/í ^sMr^^^^w^ArA/^
À5·^ï:
^"^*^^^M^^^fc
MMMMMHM «o—> ! » «» »«> w w ' » 'é» i •^.m***, i » „»im. » «»^^üU»».»^^M»w m. ! » ^»i i^|
Plate 27. Earthquake in southern Albania (Λ/=6.0, A=25 km) on March 18, 1962 recorded at Kiruna,
Sweden, at an epicentral distance of 27°. Short-period, vertical-component (uppermost trace) and medium-
period three-component (lower four traces) seismograms are displayed. This is an example of a record of
strong short-period channel (guided) waves of Lg type which traveled the continental path from southern
Albania to northern Sweden with an average velocity of about 3.4 km/s. As can be seen in the records, the
period of Lg lies between 10 and 15 s, or so. The wave train is best seen on the E-W component. Note
that the epicenter is almost due south of Kiruna. Other features exhibited in the records are the clear onset
of S and somewhat ambiguous SS arrival. P onsets are seen on vertical-components from both the short-
and medium-period seismograms. P is followed, with a 10 s delay, by another (unidentified) phase,
possibly pP.
113—114
iiBii
*S**r*f*A
'yymMpfi, /»«v*«* * . · · f
V
V VI \/ \i U V V V
V W^VV v y ^ - y ; i y ^ yv-wyli 'v y v r / " ' vv Μ ' ν ^ ν ^
Plate 28. The Michoacan, Mexico, earthquake of September 25, 1985 (Af=5.2, A=30 km) recorded at
Berkeley, California on WWSSN long-period three-component seismographs at an epicentral distance of
27e. A number of phases are identified on the seismograms. The background noise is relatively high and
thus some of the onset times are difficult to determine to the nearest second. The great circle path from
Michoacan to Berkeley runs along the continental margin between North America and the Pacific and this
leads to a complex surface wave train with surface waves arriving along both paths (oceanic and
continental). Note that the oceanic path has surface waves of a dominant period of about 25 s whereas the
continental path has a dominant period of approximately 15 s.
115
Λ
^íí~íí\'"*\ΛΛ> , Á^ÁΛ^^ >í ^ í ^ éΛ/ ^^· Λ ^ Í ^ ÁéΛΛ- ,·,
ô÷ :NW
-ºÆ
ô' f
~~Ρ ΡΡ .__ 1
II
Ä*
Plate 29. The Michoacan, Mexico, earthquake of April 30, 1986 (M=6.9, A=32 km) recorded on ultralong-
period three-component seismographs at Berkeley, California, at an epicentral distance of 27°. Note that the
horizontal components are NE or NW rather than the conventional N and E. This is because majority of
the larger earthquakes recorded at Berkeley arrive from the NW, SW and SE directions. Up on the Z, NE
and NW records corresponds to upward, northeast and northwest ground motions, respectively. Observe the
relatively simple seismograms with clear body- and surface-wave arrivals. The simplicity is largely due to
the response of the seismographs which emphasizes the long-period ground motions. As explained in the
preceding plate for this particular source-receiver position, the surface waves arrive along mixed oceanic
and continental paths. Notice that when editing this plate, the largest surface-wave amplitudes have been
cut off.
117—118
NW
[ N A / V ^ ^ ^ / ^ ^ ^ ^ —-ΛΛΛΛ/^Λ/^,ÁΛ,
s .
P ppPPP PCP
Plate 30. Ultralong-period seismograms recorded at Berkeley, California, from the October 9, 1985
earthquake south of Alaska (A/=6.5, h=30 km) at a distance of 31° northwest of Berkeley. The onsets of P,
PP, PPP and PcP are best seen on the vertical component while the S wave is clear on all three
components. Note the long-period G pulse (here G/, i.e. the wave propagating the direct route from the
hypocenter to the receiver) with a period of about 50 s, and the prominent LR waves on the Z and NW
components. The LR wave train shows an interesting and rather rare phenomenon, namely the inverse
dispersion at periods larger than 100 s. Note that the horizontal components are NE or NW rather than the
conventional N and E (for more details see the preceding plate).
119—120
-:-E
\^v\f^/*fi*jY*J/w"J^^ >TM'W^aiOAAjiu"^rVif>niW^>fy*iY'i*MV¥~«*>i" 'nfi **e+*0mf**ß^+^^00**0>m+*0mmAß***mm**0i0m^mi*EA0»*+ß
-zar
«N
~^fiWW^*f/\fl***h^^ i*^»^«^M*"M*^*¥»ii****W^ft«^^*V»>M»*^''*^**0<^^
Λ*«**^»· -rér-
~z
S
'^»*H I l M
«^^^^0«Í^**Ì^^^^^·Ì0"·«·Ì*·
PcP: :SCPPCP
* *^»
^ ^ M M A M ^ p M l M A M «
»0Nrf"^*Ä»^IP« ^ M V H M W W « M ^ « M W mnnnnwji m
-*s· LÛ-
v/
-^Λ^ν/ν^*--ΛΛ/÷\/×*ν/\/>
Plate 31. Medium size (Ëß=5.1),
shallow earthquake in the Atlantic
Ocean, off Portugal, on May 5, 1969
recorded at Umeâ, Sweden, at an
epicentral distance of 33.5°. Short-
period (upper three traces) and
medium-period (lower three traces)
three-component records are shown.
The vertical-component short-period
»N^^^^'M" ^»»* **~»-*~^~±J+^ record displays, besides the dominant
P and pP, also clear PcP and
especially pPcP. Arrival time
differences pP-P and pPcP-PcP
deduced from the record provide a
focal depth greater than normal, most
likely between 50 and 70 km. The
onset of 5 can be seen on both
:~Z horizontal-component medium-period
seismograms which also show Love
type surface waves with clear
dispersion. The Rayleigh wave train
M^»W m^vm^mt^^m+tmmnKHf**»*' can be recognized in the vertical-
component record. Note the
significantly longer period of S when
compared with that of P.
122
1
. , 1 . 1
, L, H
,—. ^
,
r
f
l
r— \ T
1 1 J L_ V 1 - : !
. y. _r j - j - f J , ,
æ"ºzz.__:__i:_:úÜiL.J.>1 é H ..MM IIL.1 . t , ,
p i — — I j l l ! 7'| M )4^f»»^HWin>i^ yewUfrf»ili.iM»#*b''wi»»»'»»i^* <ìß m»» £ £
imämmmmmmkd «11 n -rif ;n·/ r;r _ - *- - - - ·
^ ^ J ni i
1
1 > t à '! º
><! !N(cont.)
iï.-à
^^ø^Í^^,ø^^^^^,«^w^^fW
• ^ ^ . H » ^MM^^^ü^^^^m» Ç « ^ « « « ^ ^ ^^m*^^^m^^^^m^ ^ É ^ ^ Ç Μ ^ ^ ^ « Μ ^ ^ ^ ^ Μ É
\~ΛΓ^Τ-
^ ^ttM**»» 1 m *fV>WV*p>H+***A»)0*mm***0>»*>m
» i - - - LR - ,Z(cönt)
,^W^^
- v
Plate 33. Earthquake records made at Uppsala, Sweden, on three-component, medium-period seismographs.
This earthquake occurred on June 27, 1959 in the China-USSR border region at an epicentral distance of
41.5° (magnitude unknown, h=27 km). The N-S component shows a very distinct short-period Lg wave
train arriving about 14 minutes after P. Note that the epicenter is almost due east of Uppsala and hence,
Lg is best seen on the N-S component. The group traveled along a purely continental path with an average
velocity of slightly higher than 3.5 km/s. Among other features seen in the records are the clear P arrival
(vertical and E-W components) and surface waves of Rayleigh type (vertical and E-W components). There
are several arrivals in these records which have not been identified.
125—126
/A*~--<AW^
;^W^MSSS^SSSS^^W.
P*^Mw4\^'^
^V**'"">W'»* A ' , "*"~^^
}0f*+mmtt'0*-K*»
. « '
*^Ì***^1*&0Ç^·*****^04*Ú*Ú>Ç>****#* *WW>»
^^*%ÌÔ\^Λ^*Λ^^^**ÔÔ^É
P PP /i LR A M
«V^***""^A»**Vt«*«'*^
Plate 34. Long-period three-component records made at Berkeley, California, of an earthquake in the
Andreanof Islands (Λί=5.5, A=51 km) on July 1, 1982 at a distance of 42°. This is a typical set of
seismograms from a predominantly oceanic path. The records show a number of body-wave arrivals. Note
the /?P depth phase on the Z component. The 13 s interval between pP and P implies a focal depth of 48
km. The impulsive 5-wave onset is usually observed for sources below 30 km in depth in the subduction
zones. The surface waves exhibit pronounced oceanic-path dispersion characteristics. Observe the long wave
train on the vertical and E-W component that starts with a period of about 25 s and gradually decreases to
about 18 s over an interval of approximately 8 minutes.
127—
higher modes
^*^^^^^^^^^^
Plate 35. Medium-period vertical-component records of the Sinkiang earthquake (m=6.4, A=29km) on
November 13, 1965, made at Kiruna, Sweden, at an epicentral distance of 42°. This earthquake has
probably produced the strongest higher-mode surface waves ever recorded by Swedish seismographic
stations. The exhibited trace also shows a number of body-wave arrivals and an LR wave with clear
dispersion. In this particular case, the waves traveled along a purely continental propagation path.
130
:05··07^æ
•••wV^^W^^/v**^**
higher modes
I » , » » '
*^^^www^^
I
,*·*» ,«****^*·* * * \ ^ * V " ^ ^ / / \ A ' ^ * * > ^ S / ·
^ AW, V * * ^ ^ · . ^ ^ , M V ^ V . V W
AAA\W
05-.07 PP LR
r
\ l · ^++t*0*tmm wt+m,<w^l^0+J#++<»m,*m^S *******
;^'-^WVVVW
Plate 36. Seismic traces recorded at Toledo, Spain, from the underground nuclear test on November 2,
1974 on the Russian island of Novaya Zemlya (m=7.0, Af=6.2) at an epicentral distance of 42.5°. Short-
period vertical-component (upper trace) and long-period three-component (lower three traces) seismograms
are presented. These are excellent records which again emphasize the distinction between records of
earthquakes and underground nuclear explosions (see also Plate 32). We may stress that besides a strong P,
also PP and PcP are well recorded by the short-period instrument. The somewhat ambiguous phase P100P
has been interpreted as a reflection near the bottom of the lithosphère. There are two unidentified onsets
arriving 17 and 37 s after P which could be other types of PdP phases. A weak P'P' emerges about 31
minutes after P. The PKiKP, about 9 minutes after P% is hardly discernible. Long-period records show clear
high-frequency P and PP, whereas S is practically absent. Rayleigh-type surface waves with distinct
dispersion are seen on all three components. The E-W component also displays a train of higher modes.
Î I I I
20 30
A.
ß Í
^ A ¥<//Vvv-vv*^^vy*«-^«J
Ú
I1 I
P ss
^^a^v^v^WW
P! V
nent, displacement seismograms from a deep-focus earthquake in the Afghanistan-USSR border region (m=5.9, A=217
were made at KaSperské Hory, Czechoslovakia, at an epicentral distance of 43°. Displayed traces show a number of
olar P pulse and the amplitude difference between pP and sP. The epicenter is almost due east of the recording
rather weak on the N-S component. Observe the prominent long-period mantle waves of Love (N-S component) and
omponents). There is 1 minute between successive time marks at the top of the plate.
133—134
=N
the long-period E-W component seismogram marks
the calculated arrival time of S.
h*****
ËöÖöÌâÇéúúÇ**)**** ii^itMü #*«* «» I \*<$m mm^mn^nm *imw*è
i'Wyw ' I ^ ^ ^ ^ H » Î W ^ ^ H K W !
:Z
±r ^ 4 ù ^ j ) Î V » w 4 * ' ^ t-
O^MfrMjV"«»'* * > * í * » * Λ <^y>"itfw **rw*m*ém** »*tmy****,<+*+,+*>**>
f
>u
mm0*^^^1* l i
n r
*çñ**-
T T-
-^WW*· ^^^«^ ^vv*-
T T T
> ^ Mi
p- p ««***** -i r
*/^ « y i
* > Ì ^ ^ É ^ ^ *»«1Λ* ^
* ^»^ » » » » ^ í^ Λ
É ^^^ ^ × ^ ' » «
^ * ^Ml^^^^^V'WIK^II^A^^IWII
I ^í/í- *l-wi . <^í^>^»^Λ~'^^Ì^·~»*Í^»*í»í^^-'^õ»«/ííΛ/^ *^*< »w i ^»^^»ΛΛô^^ííí^- — Λ Λ Λ / W V — —
—V »
I * ^^
, Λ/ >
/>#ííÌíí^ÌÌ^Ì^"^^^*»^>^^Á*«çÃ^>/íΛ*Λ^*í^Λ '*'* ^*'**Àñ*^»'''^í''"' * é«é^>^ΛΛΛ»«*^ΛΛ^ *^*<^-*»»^>Λ/»*" ^Y^A^V^^A^W^A-A y*~>T>^sto*V~^**^^
135—136
3.0E+04
4.0E+03 ScS sS r\SS
RSNT
'" 50.1
NS ØÉÉÉÉÉÇ*-^« NS oO.l
-4.0E+03 J -3.0E+04
2.0E+03 2.0E+04
RSNT RSNT
EW 50.1 EW 50.1
-2.0E+03 -i 1 1 r -2.0E+04 n 1 é "I I >
3.0E+03q
S || ScS+SKS
5>Cb+5>KÖ
2.0E+04
ScS+SKS sS SS
RSSD 1 RSSD SA J
NS 66.9 ;
NS 66.9 3
-2.0E+04 :
-3.0E+032 2.0E+04
5.0E+03 RSSD :
93 fßty«»^»Mft«»>»»»»/f*W " ' * » ^ Wh ' EW 66.9
-5.0E+03 1 1 1 1 1 1 1 1 -2.0E+04
4
7.0E+04 P PcP 1.2E+04
RSON RSON
62.9 ||^>|ÀÉ|É».>||ÀÀ»Ì.»É|»«».É.| »H I,
62.9
-7.0E+04 -1.2E+04
3.0E+03 P PcP 1.2E + 03
LON
71.6 —ffr* LON
71.6
-3.0E+03 - 1 . 2 E + 03
3.0E+05 P PcP 7.0E + 03
GRFO GRFO Plate 41. Seismograms from a deep-focus Peru-Brazil border
92.1
-1 <»yw»»<
92.1 region earthquake on May 1, 1986 (m=6.0, A=600 km) traced
at a suite of seismograph stations that sample epicentral
distances between 50.1 and 92.6°. Short-period (left) and
-3.0E+05 -7.0E+03 long-period (right) vertical-component record sections are
7.0E+03 PPcP 1.6E + 03 P+PcP displayed. The amplitude of long-period PcP is usually of
the same order of magnitude as the coda of P. Consequently,
COL COL PcP is best distinguished from coda disturbances in records
92.6
I f f « « ! I ' flfrHW li Km j/ft,, ·.,*,;
92.6 from deep shocks, i.e. in records where depth phases appear
reliably behind PcP. As the distance increases, we note that
the delay of the depth phases, pP and sP, with respect to P
-7.OE+03: T 1 r - 1 . 6 E + 03 remain nearly unchanged. PcP, on the other hand, converges
40 80 120 160 40 80 120 160 200 240 with P. When the ray path of PcP grazes the outer core
boundary, the combination of P and PcP becomes Pc (Pdif).
Time (5) TÎme (s)
141—142
'.*W*A/*~W
!ÀÕÌÌíΛ/<·Ì^ vHV^
f5S!Jijj\^ Plate 42. Long-period three-component
5S LÛ yvwvA^A^M^ seismograms from an earthquake (Ë/=5.8,
A=27 km) south of Panama which occurred
on January 2, 1981. The recording was at
Berkeley, California, at an epicentral distance
~*^Ì~#^í**+*Ê\Í ' of 55°. The P-wave onset is clearly seen
LR along with somewhat ambiguous /?P, PcP, fT
é >
^^/»^ÍΛÁΛΛ\ÁΛΛ^^ ^ ^^VVv^^^/\Λ^ and PPP core and free-surface reflections.
These are followed by the onsets of 5, sS and
55" which show best on the horizontal
*A/C*V* ^*A^^«^\^WMAfr>*^^ i w ^ W ^ ^ components. The differential times pP-P=% s
!:««K^
and sS-S=16 S imply a focal depth of about
\ 29 km. This shallow source depth is
f WPPP
compatible with clear large amplitude Lß and
ZJ? surface waves. The surface waves have
traveled to Berkeley along a mixed
VM«^ -W^W^W continental-oceanic path, i.e. the true arrival
times of recorded surface waves do not match
either the pure oceanic or pure continental
SA/YV\AA*V\AA*^\AA*^^ travel times, they are in between.
143—144
0 1
^ÌÔÌÌÁ^^í^Ô «**·^^^*^«"
7fV·
* · *»ô»»'« i r iff * m i f » i M i
* »«n *»*" » m^pi^pi
» m*+m*mtw é
« ^Λ»«Ã».ΜΛ>"-*«* v—>
^ sS
I N First characteristics to be noted are the
absence of surface waves and large
amplitudes of P on the short-period record. It
is also interesting that pP is larger than P,
most likely due to the orientation of the
source. A number of body-wave arrivals are
displayed on the records. The arrival 24 s
after P can be explained as a reflection PdP
in the upper mantle. The onset following sP
has not been identified. Note P'P' two traces
below P (about 30 min later). On long-period
pP ,sP PPP PcS records, P phases appear as sharp pulses with
a rather short or non existent coda. Very
clear 5 and sS are exhibited on all three
S/W'^VN*^^ components.
145—146
Plate 44. Medium-period seismograms from the earthquake of October 21, 1964 (m=5.9, A=37 km) in
India-China border region, made at Uppsala, Sweden. The distance between the epicenter and Uppsala is
59°. Exhibited traces serve to illustrate records with a distinct Airy phase of the fundamental-mode
Rayleigh waves characterized by a practically constant-frequency wave train. Observe the remarkable
amplitude buildup and fall off within three swings, clearly seen in both the E-W and Z components. The
N-S component (not displayed in the plate) shows no distinct Airy phase. The vertical component also
exhibits the direct P and an unidentified phase arriving about 17 s after P. The second phase, when
interpreted as pP, provides a focal depth of 70 km.
147—148
:RE;
Plate 45. Deep-focus earthquake records made at Kiruna, Sweden, on a short-period vertical-component
(upper trace) and medium-period three-component (lower three traces) seismographs. The earthquake took
place in the Sea of Japan, at a depth of 397 km, on March 31, 1969 (m=5.7) at a distance of 63° from
Kiruna. The short-period record exhibits a clear double P9 small and large, about 4 s apart. This feature
can not be seen in any of the medium-period traces. Numerous body phases are very distinct in the
seismograms. Observe that all of them display a rather impulsive form, a typical feature for deep shocks.
Another characteristic, associated with the large focal depth, is the absence of surface waves.
Plate 46. Underground nuclear explosion on Novaya Zemlya, USSR, on October 14, 1970 (m=6.6) recorded
at Jamestown, California, at an epicentral distance of 69°. The short-period vertical-component of ground
motion is displayed. Similar to other seismograms from underground nuclear explosions, the P wave shows
a strong compression onset and the amplitude reaches its maximum within the first several swings and then
diminishes rather rapidly. The combination of a highly energetic short-period source and an epicentral
distance of 69° produces this unusual record containing multiple reflected free-surface and core phases.
Araong other features like clear PP, PPP and P'P\ we note two reflections from layers in the upper
mantle. The strong P'P' is preceded by about 20 s by a weak wavelet of P'ZOP'. The seismogram also
reveals two of the best examples of multiple reflected P waves trapped inside the Earth's liquid outer core
(Bolt, 1982). A relatively strong P4PK has been recorded 37 min and 47 s after the P wave. A weaker
P7KP is observed 1 hour and 35 s after the P onset. At an epicentral distance of 69°, P4KP travels 429°
and P7KP travels 651° around the Earth.
P1 P2 S1 S2
I I I I I I I ! ! I I I I I I I I ! ! I I I I ! I ! ! ! ! I I I I I I I I I I I I I I I I I M
Plate 47. Recording of the eruption of Mount St. Helens, Washington State, USA, on May 18, 1980, made by the GRF system in Bayern, FRG, at an
epicentral distance of 77°. The trace exhibits a computer simulated vertical-component SRO (Seismic Research Observatory) long-period seismogram
calculated from a broad-band velocity record. Seismic waves excited by the eruption have been studied in detail by Kanamori and Given (1982).
According to their analysis and interpretation, the focal mechanism of the event can be explained "... by a northward landslide followed by a lateral
blast observed at the time of the eruption. Two distinct events about 110 s apart can be identified on body-wave and short-period surface-wave
records. The first event may correspond to the earthquake which triggered the landslide and the lateral blast. The second event appears to correspond
to second large earthquake and explosion which took place about 2 minutes after the first earthquake." The two events can clearly be identified in both
the P and S wave trains.
153—154
£
relatively complex due to the size of the
source and the short-period response of the
seismograph. Note in particular the multiple
Ûj&t^^ surface reflections PP and PPP and their
corresponding depth phases. Also present are
the PKKP and PPP' core phases. The time
difference between the long path PKKP and
the direct P phases of 18 min 43 s is very
robust method for determining that the
epicentral distance is 78°. The time difference
PKKP-P is essentially independent of the
:02*38i=S=SKS: focal depth. Likewise, the time difference of
46 min 55 s between the P'P'P' and P
onsets can be used to determine the epicentral
distance. The P'P'P' phase onset is quite
clear because, in this particular case, each P'
leg travels 146° which is near the 144° caustic
for P' and thus the amplitudes are large.
Plate 49. T phase recorded on the high-frequency vertical-component seismograph at San Andreas Observatory (SAO), California, from an Ë/=8.3
earthquake which occurred in Kermadec Islands on October 20, 1986. The epicenter distance is 83° from SAO. The high-frequency seismograph has a
passband from 5-20 Hz with a magnification of 1.5 million at 5 Hz. The T phases recorded at coastal stations in California from large earthquakes
occurring in the western Pacific typically exhibit nearly monochromatic wave train at about 3 Hz which persists for 4-12 minutes. The T phase is
generally not visible on stations located more than approximately 50 km from the coastline. The largest amplitude in the T phase arrives about 88 min
50 s after the P-wave onset and its apparent velocity is 1.5 km/s (6090 s travel time over a 9140 km path).
PP PPP
I I I I I I I i I I I I ! I ! I ! I I I I I I ! I I I I I
Plate 50. The Costa Rica earthquake of April 3, 1983 (M=7.2), recorded by a vertical-component broad-band system at GRF, Bayern, FRG. The
distance is 86° and the focal depth is 44 km. There is 1 minute between successive time marks at the bottom of the plate. The P and four associated
phases that have been once or twice reflected from the Earth's surface are clearly visible in the trace. Two additional surface reflections belonging to
the S-wave group can also be identified. The complete record is, however, dominated by a fundamental-mode Rayleigh wave, developed along a
mainly oceanic propagation path. The LR wave has a long duration and a regular dispersion, within the period interval from about 30 to 15 s,
manifested by a frequency modulated quasi-sinusoidal (sweeping period) trace form. The displayed seismogram shows the ground displacement.
z
i I I I I I I
Plate 51. A shallow earthquake in northern Sumatera on November 17, 1984 (Λ/=7.3, Á=42 km), recorded at GRF, Bayern, FRG, at an epicentral
distance of 88°. Three-component, broad-band, displacement records are presented. There is 1 minute between successive time marks at the bottom of
the plate. A clear P onset is seen in the vertical trace, while the two horizontal traces reveal three later phases, namely SKS, PS and SS. Note that for
this event, the waves approach GRF almost from the east and hence SKS and SS exhibit large amplitudes on the N-S component. The distance range
between 70 and 90° is the best region to record and study SKS phases. However, the phase identification may be difficult since SKS are at these
distances often contaminated with direct S waves (see Section 4.2.3).
159—160
Plate 52. Records from a deep-focus earthquake in Celebes Sea (m=6.2, A=336 km) of June 11, 1972,
made by medium-period seismographs at Kiruna, Sweden, at an epicentral distance of 91.5°. A double P
onset, small and large, some 3 s apart, is discernible on the vertical component. The trace also shows a
sharp pP arrival about 75 s after P providing a focal depth of about 320 km. pP is followed by clear free-
surface reflection PP. Among other features can be seen the long-period G wave train especially well
developed on the N-S componenL The wave has traveled along a mixed oceanic-continental path of about
10,160 km with an average velocity of 4.52 km/s and appears nearly impulsive with a period of
approximately 30 s. Note that in this record successive traces go from bottom to top.
161
Plate 53. The Philippine Islands deep-focus earthquake of November 20, 1984 (m=6.4, A=215 km) recorded
at Mt. Hamilton, California, on a short-period, vertical-component seismograph. This earthquake occurred at
a distance of 105° from Mt. Hamilton. Note the numerous body phases including free-surface reflections.
Note also that the dominant period of the S phases is around 4 s and longer, while for the P phases it is 2
s and less. The depth of focus for this event can be found from several onset-time differences such as pP-
P, pPP-PP, sSKS-SKS and SKS-P. Of these, pP-P is most commonly used but SKS-P is of interest because
it does not have a strong distance dependence. The P400P phase, which reflects off the underside of the
400 km discontinuity is sometimes observed at Mt. Hamilton for Southern Pacific sources. The time
difference PKKP-P of 15 minutes 43 s is used to determine the epicentral distance of 105°.
163—164
-4.0E + 0 3 J -8.0E + 03
5.0E+02.PKIKP + PKP2
i.OE+03: PKP1
PKP2 PKP1
SCP 3PK\KP
^i^^H^f·^^
SCP
156.9 156.9
-1.0E+034 -5.0E+024
6.0E + 02.PKIKP 9.0E + 02
PKP2 PKP2 Plate 55. Seismograms from a Mid-Indian-Rise
JAS JAS earthquake on May 16, 1985 (Ëß=6.0, /é=10 km)
162.5 Μ^φ^ΗΗ^ 162.5
made at a suite of seismograph stations that sample
the epicentral distance between 134.3° and 173.2°.
Short-period (left column) and long-period (right
column) vertical-component records are displayed.
-6.0E+02 D -9.0E + 02 J Station codes, epicentral distances and amplitude
scalings are given to the left of each trace. This
3.0E + 033 8.0E + 03q example illustrates the effect of the caustic near 144°.
PKP2
RSSD PKIKP PKP2 RSSD PWKP Note the large concentration of energy on records
from station RSNT at a distance of 145.6°. Besides
164.9 ÌøÇ^Ì^ 164.9 large amplitudes in both the short- and long-period
range, we also observe almost simultaneous arrival of
several waves which makes the separation into
-3.0E + 0 3 j -8.0E + 03 d individual PKP branches impossible. First at the
distance of 156.9° (stations SCP), the three onsets,
1.4E4-05, PWKP 9.0E + 03q corresponding to PKIKP, PKP1 and PKP2 arrivals,
PKP2 PKP2
ANMO ANMO become discernible. PKPl dominates the records at
distances just beyond the caustic but diminishes
173.2 3— m^^r^kAi^^fy4m 173.2 rather rapidly as the distance increases and vanishes
from the seismogram at a distance of about 160°.
Starting from the record made at SCP at a distance
-1.4E + 05 + -9.0E + 03 of 156.9°, PKP2 arrivals dominate the short-period
40 80 120 160 2( 40 80 120 160 200 seismograms.
REFERENCES
Adams, R.D., 1979. T-phase recordings at Rarotonga from underground nuclear explosions. Geophys. J.R.
astr. Soc., 58: 361-369.
Bolt, B.A., 1982. Inside the Earth. Freeman and Company, San Francisco, 191 pp.
Bullen, K.E., 1975. The Earth's Density. Chapman and Hall, 420 pp.
Bullen, K.E. and Bolt, B.A., 1985. An Introduction to the Theory of Seismology. Cambridge University
Press, Fourth edition, 499 pp.
Bungum, H., Vaage, S. and Husebye, E.S., 1982. The Mel0y earthquake sequence, Northern Norway;
source parameters and their scaling patterns. Bull. Seismol. Soc. Am., 72: 197-206.
Bath, M., 1947. Travel times of the principle earthquake waves for Uppsala. Bull. Geol. Instit., Uppsala,
32: 105-129.
Bath, M., 1967. Observations of teleseismic Pn phases. Pure Appl. Geophys., (PAGEOPH), 66: 30-36.
Bath, M. and Shahidi, M., 1974. Γ-phases from Atlantic earthquakes. Pure Appl. Geophys., (PAGEOPH),
92: 74-114.
Bath, M., 1979a. Introduction to Seismology. Birkhauser, 428 pp.
Bath, M., 1979b. Earthquakes in Sweden 1951-1976. Sveriges Geol. Und., C750, Arsbok 72, 79 pp.
Bath, M., 1983. Earthquake data analysis: An example from Sweden. Earth-Science Rev. 19: 181-303.
Frohlich, C , 1987. Kiyoo Wadati and early research on deep focus earthquakes: Introduction to special
section on deep and intermediate focus earthquakes. J. Geophys. Res., 92: 13,777-13,788.
Frohlich, C , 1989. Deep earthquakes. Scientific American, January 1989: 32-39.
Gutenberg, B. and Richter, C.F., 1938. Depth and geographical distribution of deep-focus earthquakes.
Geol. Soc. Am. Bull., 49: 249-288.
Herrin, E., Arnold, E.P., Bolt, B.A., Clawson, G.E., Engdahl, E.R., Freedman, H.W., Gordon, D.W., Hales,
A.L., Lobdell, J.L., Nuttli, O., Romney, C , Taggart, J. and Tucker, W., 1968. 1968 Seismological
Tables for Ρ Phases. Bull. Seismol. Soc. Am., 58: 1193-1352.
Jeffreys, H. and Bullen, K.E., 1967. Seismological Tables. Brit. Assoc. Adv. Sci., 50 pp.
Kanamori, H. and Given, J.W., 1982. Analysis of long-period seismic waves excited by May 18, 1980,
eruption of Mount St. Helens - a terrestrial monopole? J. Geophys. Res., 87: 5422-5432.
Kind, R., 1979. Observations of sPn from Swabian Alp earthquakes at the GRF array. J. Geophys., 45:
337-340.
Lee, W.H.K., 1987. Observational Seismology. Encyclopedia of Physical Science and Technology, Vol 12:
491-518. Academic Press.
Minakami, T., 1959a. The study of eruptions and earthquakes originating from volcanoes: Part 1. Bull.
Volcanol. Soc. Jpn., 4: 104-114 (in Japanese).
Minakami, T., 1959b. The study of eruptions and earthquakes originating from volcanoes: Part 2. Bull.
Volcanol. Soc. Jpn., 4: 115-130 (in Japanese).
Minakami, T., 1960. The study of eruptions and earthquakes originating from volcanoes: Part 3. Bull.
Volcanol. Soc. Jpn., 4: 133-151 (in Japanese).
Neumann, F., 1951. Principles Underlying the Interpretation of Seismograms. U.S. Department of
Commerce, Coast and Geodetic Survey. Special Publication No. 254, 41 pp.
Payo, G., 1986. Introduccion al Analisis de Sismogramas. Instituto Geografico Nacional, Madrid, 125 pp.
Pilgrim, L., 1913. Die Berechnung der Laufzeiten eines Erdstosses mit Beriicksichtigung der Herdtiefen,
gestutzt auf neuere Beobachtungen. Gerlands Beitr. z. Geophys., 12: 363-483.
Press, F. and Siever, 1982. Earth. Freeman and Co., Third edition, 613 pp.
Seidl, 1986. Analysis of Grafenberg broadband seismograms. In: B. Buttkus (Ed.), Ten Years of the
Grafenberg Array. Geol. Jahrbuch, Reihe E, Geophysik, Heft 35, 135 pp.
Shida, T., 1937. Thank-you address at the dedication ceremony of Beppu Geophysical Laboratory (in
Japanese). Chikyu Butsuri Geophysics, 1: 1-5.
Simon, R.B., 1968. Earthquake Interpretations. Colorado School of Mines, 99 pp.
Tazieff, H. and Sabroux, J.C., (Eds.), 1983. Forecasting Volcanic Events. Developments in Volcanology 1,
Elsevier, 635 pp.
Tokarev, P.I., 1983. Experience in predicting volcanic eruptions in the USSR. In: Tazieff, H. and Sabroux,
J.C. (Eds), Forecasting Volcanic Events. Developments in Volcanology 1, Elsevier, 635 pp.
170
Turner, H.H., 1922. On the arrival of earthquake waves at the antipodes, and on the measurement of the
focal depth of an earthquake. Mon. Not. R. Astron. Soc., Geophys. Suppl., 1: 1-13.
Wadati, K., 1927. Existence and study of deep earthquakes (in Japanese). J. Meteorol. Soc. Jpn., Ser. 2, 5:
119-145.
Wielandt, E., 1983. Design principles of electronic inertial seismometers. In: Kanamori, H. and Boschi, E.
(Eds.), Earthquakes, Observations, Theory and Interpretation. North-Holland, 608 pp.
Willmore, P.L., 1979. Manual of Seismological Observatory Practice. Report SE-20, World Data Center A
for Solid Earth Geophysics, NOAA, Boulder.
171
SUBJECT INDEX
- - see also Microseisms Pg, 20-25, 28, 51, 53, 54, 66, 70, 4, 5, 12, 21
Olivine, 40 -, multiple, 6
One-sided onset, 10
see also Unipolar onset Pg2,6
Origin time, 48 Phase, 19, 49
-, Airy, 43
p , 30 -, depth, 24, 30, 31, 37, 38, 49
pP, 31, 38, 49, 70, 27, 3 1 , 34, 35, 37, 41-45, - nomenclature, 69
48, 52, 53 -, stopping, 65
pPc, 54 - transition, 40
pPcP, 3 1 , 37, 48 PKIIKP, 35, 71
pPcS, 37 PKiKP (PKP ), 35, 36, 7 1 , 36
CD
49-52, 66, 70, 3 , 7-9, 11, 14, 15, 23-39, PKP2 (PKPu), 34-36, 7 1 , 55
41-52 PKS, 33, 35, 70
-, double, 65, 45, 52 Plane of weakness, 4
PI, 65, 47 see also Fault plane
P2, 65, 47 Plate deformation, 4
P100P, 36 -, major, 3, 4
P400P, 3 1 , 53 -, secondary, 3
P4PK, 36, 46 - tectonics, 3, 40
P650P, 31 PmKP, 30, 36, 37, 71
P7PK, 36, 46 PmP, 20, 23, 53, 70, 4
Λ 20, 22, 23, 70, 1 Pn, 21, 23-25, 28, 29, 51, 53, 54, 70, 4, 10, 12,
Ρ/, 2 13
P2, 2 Poisson ratio, 41
P*, 22, 23, 28, 53, 54, 70 Polarization, 15, 41
P \ 33 PP, 29-31, 47, 49, 51, 70, 26, 29, 30, 33-37, 39,
- see also PKP 42, 46, 48, 50, 52-54
P ' 8 0 P \ 46 -, early, 32
P ' 6 5 0 P \ 36, 46 see also Precursors to P P
P", 54 PPP, 29, 70, 30, 42, 43, 46, 48, 50
-, see also PKIKP PPS, 29, 70, 50
Pacific plate, 5 PS, 29, 50, 51
Particle motion, 41 PSP, 29, 70
Pb, 22, 70 PSS, 29, 70
Pc, 32, 35, 47, 70, 4 1 , 53, 54 Precursor (Forerunner), 36
- see also Ρ diffracted - to PP, 32
Ρ coda, 17, 19 Propagation path, 13
PcP, 30, 33, 38, 47, 49, 51, 70, 26, 30, 3 1 , 36, - -, continental, 15, 22, 25, 42, 43, 22, 27-29,
41, 42, 46 33, 39
PcS, 30, 33, 49, 70, 34, 37, 43 - -, mixed, 23, 24, 29, 42, 52
Ρ diffracted (Pc, Pdif, Pdiff), 30, 32, 33, 70, 41 - -, oceanic, 15, 42, 43, 28, 29, 34, 50
PdP, 32, 36, 70, 36, 43, 46 PR1, 11
P'dP', 36, 71 PR2, 11
Pendulum, 59 Principle earthquake, 5
- period, 60 - -, see also Main shock
- principle, 55 Ρ (Primary) wave, 14, 16-20
174
- - see also Compressional wave 52, 4
P'P', 36, 43, 46, 48 S*, 22, 23, 28, 53, 54, 70
P'P'P', 36, 48 San Andreas Observatory (SAO), 77, 49
Sb, 22, 70
Quarry blast, 2 1 Sc (Sdif, Sdiff), 32, 70
Querwelle, 14 - see also S diffracted
Scattering, 16, 36
R, 44, 7 1 , 14, 37 ScP, 30, 33, 49, 70
Rl, 45, 71 ScS, 30, 33, 38, 39, 70, 40, 45 48
R2, R3 etc., 45 S diffracted, 32
Radio time signal, 57 - see also Sc (Sdif, Sdiff)
Ray, 19-21, 29, 34, 35 Secondary discontinuity, 31
- path graph, 4 Seismic Research Observatory (SRO), 47
Rayleigh wave, 14, 15, 41-44, 49, 7 1 , 12, 24-26, Seismic source, 7, 8
3 1 , 33, 36, 44, 50 - -, controlled, 7
- - see also LR - -, man-made, 7
- -, omnidirectional, 21 - -, natural, 7
- -, short-period, 5, 21 Seismogram (Record), 55
Recording, 57 -, analog, 65, 68, 69
-, broad-band, 10, 12 -, broad-band, 26, 64, 65, 25, 37, 47, 51
-, digital, 68 -, digital, 65, 66, 68, 69
-, electromagnetic, 57, 58 -, displacement, 64, 10, 12, 25, 37, 50, 51
-, electrostatic, 57 -, long-period, 37, 39, 64, 65, 67, 22-24, 32, 34,
-, galvanometric, 58 36, 38-43, 47, 55
-, mechanical, 57, 58 -, medium-period, 27, 3 1 , 35, 44, 45
-, mechanical-optical, 57, 58 -, short-period, 27, 37, 39, 64, 65, 6, 8, 9, 13,
- on photographic paper, 57, 58, 65, 67 15-18, 27, 3 1 , 32, 36, 40, 4 1 , 43, 45, 46, 55
Reservoir induced seismicity, 7, 8 -, ultralong-period, 30
Resonance, 60 -, velocity, 10, 12, 47
Response characteristic, 60-62 Seismograph, 55-58
see also Magnification or Response -, broad-band, 18, 6 1 , 62, 69, 12, 50
characteristic -, high frequency, 49
Rg, 24, 25, 28, 53, 70, 5, 21 -, long-period, 17, 18, 37, 6 1 , 62, 66, 69, 26,
Ridge, 3, 4, 6 28, 54
Ripple charge blasting, 2 1 -, Mainka, 57, 58
Rockburst, 24, 72, 73, 5 -, medium-period, 69, 26, 33
Rupture duration, 13 -, Milne-Shaw, 57, 58, 62
- propagation, 13 -, ocean-bottom, 7
-, portable, 18
s, 30 -, short-period, 27, 37, 61, 62, 69, 1, 3 , 14,
sP, 31, 49, 70, 37, 4 1 , 43, 53 19-21, 26, 48, 53
sPcP, 37 -, ultralong-period, 29
sPn, 24, 70, 10 -, Wiechert, 57, 58, 6 1 , 62
sPP, 31, 70 -, Wood-Anderson, 57, 58, 61
sPS, 31, 70, 53 Seismological Observatory, 55, 61
sS, 31, 37, 38, 70, 40, 42, 43, 45 Seismometer, 55, 58-60, 65
sSKS, 53 -, Benioff, 55, 58, 6 1 , 62, 14
sSP, 31 -, deformation see Strain seismometer
sSS, 31 -, electrostatic, 58
S, 11, 19, 2 1 , 23, 25-32, 36-41, 43, 47, 49, 50, -, Galitzin, 57, 62
64, 70, 3 , 7-9, 11, 14, 15, 22-31, 34, 35, -, Grenet-Coulomb, 56, 58, 6 1 , 62
38-40, 42, 43, 45, 47, 48 -, Kirnos, 61, 62, 64
51, 47 -, moving-coil, 56
52, 47 -, ocean-bottom (OBS), 26
. ^ 2 0 , 22, 23, 70, 1 -, pendulum, 56, 59
SI, 4 -, piezoelectric, 57
175
Muotatal, 77 78, 5, 22, 24, 26, 27, 31-33, 35, 38, 39, 44,
Mururoa Atoll, 14 45, 52
Switzerland, 75-78, 4
Nagano, 77
Nevada, 76, 77, 20 Tacana volcano, 76, 15
New Guinea, 73 Tamaro, 78
New Hebrides Islands, 38 Tangshan, 5
New Ireland, 72, 54 Tenessee, 77
New Madrid, 5 Thailand, 37, 77
New Mexico, 77 Tokai District, 78, 7
New York, 77 Toledo, 78, 23, 36, 43, 54
New Zealand, 74, 77, 8 Tonga Islands, 75, 14
Niue, 77, 14 Tonga-Kermadec Islands, 38
North America, 28 Tuamotu Archipelago, 75
Northwest Territories, 77
Norwegian Sea, 27, 72, 73, 25 Uddeholm, 78, 5
Noto Peninsula, 9 Umbria, 23
Novaya Zemlya, 72, 36, 46 Umea, 18, 27, 39, 78, 31
United States, 57, 73, 77
Ontario, 77 Uppsala, 17, 18, 26, 39, 45, 78, 5, 22, 24, 26,
32, 33, 38, 39, 44
Pacaya volcano, 76, 18 USSR, 72-74, 32, 33, 37, 43, 46
Pacific, 49, 28, 49, 53
Panama, 73, 42 Washington State, 73, 77, 47
Papua, 54
Pennsylvania, 77 Xinjiang, 76, 38
Perth, 5
Peru, 75, 41 Yellowknife, 77
Philippine Islands, 53
Poas volcano, 76, 16
Portugal, 31
Priest Mountain, 20, 21
Rarotonga, 77, 14
Red Lake, 77
Rincon de la Vieja volcano, 75, 77, 19
Romania, 64, 65, 72
Schleitheim, 77
Sea of Japan, 62, 72, 73, 45
Sea of Okchotsk, 38, 39, 74, 40
Semipalatisk, 32
Shizuoka, 77
Sibinal, 77, 15
Siberia, 76, 39
Sierra Nevada, 21
Sierre, 4
Sinkiang, 72, 35
South America, 38
South Dacota, 77
South Island, 74, 77
Spain, 40, 78, 23, 36, 43, 54
State College, 77
Sumatera, 74, 51
Swabian Jura, 10
Sweden, 17, 26, 27, 45, 53, 68, 72, 73, 75, 77,