F U R T H E R TITLES IN T H I S SERIES

VOLUMES 1-5 AND 10-12 ARE OUT OF PRINT

6 L. CIVETTA, P. GASPARINI, G. LUONGOandA. RAPOLLA (Editors)

PHYSICAL VOLCANOLOGY

7 M. BATH
SPECTRAL ANALYSIS IN GEOPHYSICS

8 O. KULHANEK
INTRODUCTION TO DIGITAL FILTERING IN GEOPHYSICS

9 T. RIKITAKE
EARTHQUAKE PREDICTION

13 V.C. DRAGOMIR, D.N. GHITAU, M.S. MIHAILESCUand

M.G. ROTARU
THEORY OF THE EARTH'S SHAPE

14A A.J. BERKHOUT

SEISMIC MIGRATION
Imaging of acoustic energy by wave field extrapolation
B. Practical aspects

15 E. BISZTRICSANYand GY. SZEIDOVITZ (Editors)

PROCEEDINGS OF THE SEVENTEENTH ASSEMBLY OF THE
EUROPEAN SEISMOLOGICAL COMMISION

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

For the IASPEI/Unesco Working Group on

I A S Ρ ΕI Manual of Seismogram Interpretation

ELSEVIER
Amsterdam — Oxford — New York — Tokyo
ELSEVIER SCIENCE B.V.
Sara Burgerhartstraat 25
P.O. Box 211, 1000 AE Amsterdam, The Netherlands

First edition: 1990

Second impression: 1997

ISBN: 0 444 88375 4

© 1990 ELSEVIER SCIENCE B.V. All rights reserved.

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.

This book is printed on acid-free paper

Printed in The Netherlands

 ν

/ remember when our island was shaken by an earthquake

some years ago, there was an impudent mountebank who
sold pills which as he told to the country people were very
good against an earthquake,
"The Tatler", Joseph Addison 1672-1719
VI

PREFACE

Numerous manuals for seismogram interpretation and analysis have been

circulating among seismologists over the past several decades. Many of them,
often issued as in-house handbooks guiding analysts at a particular
seismographic observatory, have been of rather local or regional importance.
Some have gained world-wide recognition and a few are still employed in
routine work at observatories around the world as well as for training
activities of various kinds. For instance, the special publication Principles
Underlying the Interpretation of Seismograms has long been used as a
standard reference book. It was written in 1951 by F. Neumann, a
geophysicist affiliated with the U.S. Coast and Geodetic Survey, and
necessarily describes routines commonly employed at that organisation. In
1968, R. B. Simon of the Colorado School of Mines wrote Earthquake
Interpretations. She collected an impressive suite of records made at
seismographic stations Bergen Park (Colorado) and Palisades (New York) and
presented the corresponding phase identifications and commentaries. In 1986,
G. Payo published his excellent Introduccion al Analysis de Sismogramas. His
work, written in Spanish, comprises a short introduction and 87 plates with
seismograms from the Toledo seismographic station in southern Spain.
Extensive commentaries are included.

All the above publications make use of seismic records produced by

standard analog narrow-band seismographs and illustrate rather well the
developments since the early 1950's. It is obvious that any similar book can
hardly be considered as a perfect final product, simply because our highly
fragmental knowledge of the Earth's interior, the available instrumentation,
associated interpretation and analysis techniques and theory are continuously
improving. Hence, a frequent updating of existing interpretation codes and
routines is not only welcomed but is in fact a prerequisite for modern
seismogram interpretation, as well as for research.

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.

In the course of this book, I shall strictly distinguish between terms

seismogram interpretation and seismogram analysis. The former, being the
content of the book, is devoted to the art of identification of various seismic
wave types appearing on seismograms, including the recognition of the waves
 VII

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.

The present publication arose as a joint IASPEI/Unesco venture. Both

bodies have shared an interest in editing a new comprehensive manual on
seismogram interpretation. Discussions concerning the manual had already
started at the London, Ontario, IASPEI Assembly in 1981. However, the first
definite steps to tackle the problems were only made at the Tokyo IASPEI
Assembly in 1985. During this meeting, a IASPEI/Unesco Working Group
(WG) on Manual of Seismogram Interpretation was established. Two business
meetings were held and specified the general requirements of the manual.
Initially, the WG consisted of G.L. Choy, V.I. Gorbunova, M. Hashizume
(Unesco representative), O. Kulhanek (chairman), D. Mayer-Rosa, L.
Ruprechtova, M.E. Reyners, D. Seidl, R.A. Uhrhammer and M. Yamamoto.
The first selection of material to be included in the manual was made at the
subsequent EGS/ESC Assembly in Kiel, FRG, in 1986. G. Payo became a
new member of the WG while R.E. Reyners resigned. The final lay-out and
definite selection of seismograms were considered at several business
meetings of the WG during the IUGG General Assembly in Vancouver,
1987.

This manual would not have been written without the support, continuous
interest and encouragement of Unesco. Invaluable indeed are the seismogram
VIII

examples with commentaries received from the following WG members and

others:
G.L. Choy Plates 40, 4 1 , 55
D. Mayer-Rosa Plate 4
E. Molina Plates 15, 18
G.Payo Plates 23, 36, 43, 54
A. PleSinger Plates 2, 37
J.M. Protti Plates 16, 17, 19
M.E. Reyners Plates 8, 14
D. Seidl Plates 10, 12, 25, 47, 50, 51
R.A. Uhrhammer Plates 20, 21, 28-30, 34, 42, 46, 48, 49, 53
M. Yamamoto Plates 1, 3, 6, 7, 9, 11, 13, 15-17, 19
The remainder, Plates 5, 22, 24, 26, 27, 31-33, 35, 38, 39, 44, 45, 52, were
compiled by the author.

I would like to thank my collegues A.J. Anderson and J. Henderson who

carefully read and commented on various parts of the manuscript. I
appreciate very much the long hours of reading put in by R.D. Adams and
G.L. Choy who provided me with thorough reviews of the entire manuscript.
They were my most useful critics and made many good suggestions which
have been incorporated in the text. Illustrations supplementing the text were
prepared by H. Nilsson. Anne-Marie Olsson and Siv Petterson patiently and
carefully typed countless revisions of the manuscript.

Seismological Section Ota Kulhanek

Uppsala University May, 1989
Uppsala, Sweden
 1

Chapter 1

INTRODUCTION

Earthquakes are mighty manifestations of sudden releases of strain energy

accumulated during extensive time intervals in the upper part of the Earth.
They radiate seismic waves of various types which propagate from the
earthquake in all directions through the Earth's interior and are recorded at
large distances by sensitive instruments placed on or near the Earth's surface.
The appearance of an actual seismogram reflects the combined effects of the
source, the propagation path, the characteristics of the recording instrument
and the ambient noise due to the specific conditions at the particular
recording site. To understand the rather complicated nature of seismogram
traces requires knowledge of seismic source physics, internal structure of the
Earth and seismic wave propagation but mainly it requires long experience
based upon daily inspection of actual seismic records. A trained seismogram
analyst frequently reveals and in many cases correctly interprets record
features and details invisible to other workers, well in agreement with Emil
Wiechert's "... jede Zacke, jede Zunge zu erklaren.." (from German meaning
"... every jerk, every wiggle should be explained..."). It is well known that
veteran seismogram interpreters who have been associated for a long time
with a given seismographic station often develop miraculous abilities to
recognize a characteristic appearance of seismograms from earthquakes and
other seismic phenomena occurring practically at any part of the world.

In the following sections, a brief description of fundamental physical

phenomena affecting the seismogram appearance is presented. The second
part of this book contains a number of plates with actual seismogram
examples and corresponding interpretations. It should be emphasized that
current seismogram interpretation together with subsequent analysis is not
limited to determination of basic earthquake source parameters such as
location, time of occurrence, focal depth and magnitude. Reflected and
refracted seismic waves are used to test hypotheses concerning the Earth's
internal structure as well as to discover, locate and describe new
discontinuities and other features deep inside the Earth. Advanced techniques
are used to study in detail the physical processes and stress distribution in,
and close to, earthquake foci. Thousands of earthquakes are utilized in studies
of geological movements and in earthquake risk mapping. Seismic records
also provide the first-hand information when identifying various precursory
phenomena associated with large earthquakes and are, therefore, of great
value for earthquake prediction. There is no doubt that the overwhelming part
2

of contemporary knowledge of the internal structure of the Earth and its

dynamics has been inferred from seismological studies employing
seismograms as primary data. We are aware of the disasters frequently
caused by earthquakes, but are powerless to interrupt geological processes
taking part inside our planet. However, by learning about these processes,
and there is still much to learn, we increase our chance of mitigating the
earthquake threat.

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

EARTHQUAKES, WHY AND WHERE DO THEY OCCUR?

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.

In plate tectonics, the uppermost part of the Earth is considered to be

divided into two layers with different deformation properties. The upper rigid
layer, called the lithosphere, is about 100 km thick below the continents,
about 50 km under the oceans, and consists of crust and rigid upper-mantle
rocks. The lower layer, called the asthenosphere, extends down to about 700
km depth and in it the mantle rocks are less brittle, i.e. more deformable
when compared with the lithosphere. The rigid lithospheric shell is broken
into, say, a dozen irregularly shaped major plates (not coinciding with
continents) and a large number of minor or secondary plates. Figure 1
displays the division of the Earth's surface into major rigid plates. The
lithospheric plates are not stationary; on the contrary, they "float" in a
complex pattern, with a velocity of some 2-10 cm/year on the soft rocks of
the underlying asthenosphere like rafts on a lake. The idea of wandering
plates was originally proposed in 1912 by a German scientist, A. Wegener.

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).

It becomes obvious, even from this highly simplified description of

plate-tectonic processes, that a continuous accumulation of stresses takes
place in rocks along plate margins at various depths in the Earth. Plate
deformation (bending, compression, tension), friction between colliding plates,
high temperature gradients within subducting slabs and so forth contribute
significantly to the build-up of stress. When the stress exceeds the elastic
strength of the rock, the rock fractures along a plane of weakness (fault
plane) and an earthquake is born. The rock is displaced into its new position
while the strain energy is totally or partially released. Rock fracturing usually
starts from a point (focus, hypocenter) close to one edge of the fault plane
and propagates along the plane with a typical velocity of some 3 km/s. The
vertical projection of the hypocenter onto the Earth's surface is called the
epicenter. After large shallow earthquakes, the observed surface dislocations
can amount to several meters and the fault length can exceed hundreds of
kilometers.

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

others. There is a high correlation between this geographic distribution of

epicenters (interplate earthquakes) and plate margins. A large part (80%) of
the seismic energy released by all earthquakes is released along the margins
of the Pacific plate, also called the Circum-Pacific belt. A high energy
concentration can also be seen along the Mid-Atlantic ridge and the Alpide
(Asiatic-European) belt extending from the Azores to southeastern Asia.
Sporadically, earthquakes also occur at rather large distances from the
respective plate margins, such as the large earthquakes in central USA (New
Madrid, 1812; Charleston, 1886), southwestern Australia (Perth, 1968),
northeastern continental China (Tangshan, 1976) etc. These, so called
intraplate earthquakes, show a diffuse geographical distribution and their
origin is still poorly understood. These earthquakes can be large and because
of their unexpectedness and infrequency can cause major disasters.

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).

Major earthquakes are often followed by a number of aftershocks, i.e. lesser

earthquakes that follow the main shock or principal earthquake and
originate near the focus of the main shock. Aftershock sequences may
continue for several days, weeks, months and even years. For instance, after
the great Kamchatka earthquake on March 4, 1952 (M=8.6), the activity in
the area remained higher than normal for 3-4 years (B&th, 1979a). Generally,
the frequency occurrence of aftershocks (which in some cases may be of the
order of a hundred or more per day) decreases with increasing time.
6

Sometimes, the main shock is preceded by foreshocks, which are usually

smaller quakes occurring shortly, several days or weeks, before and near the
main shock. It should be emphasized that significant deviations from this
pattern of energy release i.e. foreshocks - main shock - aftershocks, are quite
common. Weak earthquakes, with magnitudes 3 or less, are often called
microearthquakes.

Still another mode of energy release is manifested by earthquake swarms.

In these, a large number of earthquakes takes place within a limited area
over time periods from a week or so, to several months. An earthquake
swarm shows no pronounced main shock and the frequency of shocks
gradually increases until a maximum is reached and then the activity
gradually dies out. As an example, we can mention the swarm activity in
Meloy, northern Norway, which began in November 1978 and the high
activity lasted until the end of January 1979. During this time interval, more
than 10,000 tremors (M < 3.2) were recorded by nearby stations. At its
maximum activity, the harvest for one single day was more than 800
microearthquakes (Bungum et. al, 1982). However, cases of swarms are
known where the maximum frequency was as high as several thousands of
small quakes per day. Swarms are also common in volcanic regions, as will
be discussed later.

It is worth mentioning that the strain energy accumulation is rather slow,

taking months, years and even decades (depending on the size of the
earthquake, i.e. on its magnitude) before the level of rock failure is reached.
On the other hand, the energy release takes only seconds or a fraction of a
second for small earthquakes and between one and a few minutes for the
largest shocks. Modern research reveals that a smooth continuous rupture
over the entire active fault plane is an exception rather than a rule. The usual
case, especially for large shocks, is an irregular rupture process through a
series of partial dislocations. Such a multiple rupturing, of course, heavily
increases the complexity of corresponding seismograms.

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

Earthquakes described above, i.e. those caused by the sudden release of

accumulated strain energy, due to interaction of two or more lithospheric
plates, are distinguished as tectonic earthquakes (in Greek tecton means a
builder). Other categories of earthquakes are volcanic earthquakes and
implosions or collapse earthquakes. Volcanic earthquakes, discovered by L.
Palmieri at the Vesuvius Observatory in 1855, are caused by sudden opening
of channels in crustal rocks, rapid changes of motion of magma, excessive
accumulation of gas pressure in the crust, roof collapses of subterranean
channels emptied of magma, etc. According to Minakami (1959a, b, 1960) or
Tazieff and Sabroux (1983), volcanic earthquakes are classified into three
groups: Α-type earthquakes with foci between 1 and 10 km deep; B-type
earthquakes with foci at depths of 1 km or less; and explosion-type
earthquakes taking place at the very surface of the Earth. Another
categorization of volcanic earthquakes can be found e.g. in Tokarev (1983).
Close to active volcanoes, we also frequently detect so called volcanic
tremor which is due to long-duration, more or less continuous, volcanic
vibrations. Whereas volcanic earthquakes are clearly individual events
separated in time from each other, volcanic tremor shows rather spasmodic or
harmonic behaviour. It is associated with flow of underground magma,
oscillations in magma reservoirs, explosions of volcanic gases, etc.

Many tectonic earthquakes and some volcanic earthquakes are monitored at

large distances. Collapse earthquakes, which are due to collapses of
subsurface cavities, are generally only of local significance and are practically
always rather small.

Tectonic, volcanic and collapse earthquakes together with oceanic

microseisms, discussed in Section 4.1, belong to the category of natural
seismic sources. There is also a variety of man-made seismic sources such
as industrial or military explosions and various types of cultural noises
(traffic, industry, construction works), which are examples of controlled
seismic sources where place, time of occurrence and source intensity are
determined in advance or are, at least, highly predictable. Other types of
man-made seismic sources are induced or triggered events. Generally
speaking, there are two plausible mechanisms available to explain triggered
events. Firstly, by changes in local elastic stresses (loading, unloading) caused
by removal of large volumes of rocks as in mining and quarrying operations
and by reservoir impounding. Secondly, by an increase of pore and fracture
pressure, e.g. due to fluid injection, which in its turn decreases the rock
strength (it acts as a lubricant) and may thus give rise to an increase of local
seismicity. The best documented cases of triggered seismicity are those
associated with the injection of fluids under high pressure into crustal rocks.
Physical explanations of reservoir induced seismicity are still imperfectly
understood but the impounding of large reservoirs can obviously affect both
the local elastic stress as well as the fluid pressure. It has to be emphasized
8

that, irrespective of various trigger mechanisms, reservoir induced events

generally also relieve tectonic stresses built up over very long time periods,
and hence are about to occur anyhow. To some degree, hypocentral locations
of induced earthquakes are predictable. However, their size and time of
occurrence are not. Many man-made events such as underground nuclear
explosions and some of the reservoir-induced earthquakes, are large enough
to be received teleseismically at seismograph stations throughout the world. A
brief overview of major seismic source types is sketched in Fig. 3.

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

Fig. 3. Major categories of seismic sources

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

STRUCTURE OF THE EARTH'S INTERIOR

Even the most elementary seismic waves recorded by a seismograph station

cannot be described and discussed without having first a working model of
the Earth's interior through which the waves travel. For seismological
purposes, it is convenient to assume the Earth to be constituted of crust,
mantle and core (Fig. 4). This major division was established from the
analysis of recorded seismic waves in the beginning of this century and still
provides a reasonable working model. The mantle-crust as well as the
core-mantle boundary are distinct discontinuities in seismic-wave velocities
and efficient reflectors/refractors of the incident seismic energy.

3.1 CRUST

The mantle-crust boundary, generally called the Mohorovicic discontinuity

(often abbreviated as Μ or Moho), separates rocks at the base of the crust
with compressional-wave velocities of about 6.5 km/s, from the underlying
mantle rocks where compressional-wave velocities are about 8 km/s (Fig. 5).
The average thickness of the crust varies from about 25 to 40 km below the
continents but may be as large as 60 to 70 km under high mountains. Under
the deep ocean, the crust is much thinner, only about 5 km. In studies of
nearby earthquakes, epicentral distance less than 1000 km, we often assume a
crust consisting of two horizontal layers of approximately the same thickness,
separated by the Conrad discontinuity. The upper layer represents granitic
rocks, whereas the lower layer consists of basaltic rocks. For a typical crustal
model under the deep ocean we usually omit the granitic layer.

3.2 MANTLE

The Earth's mantle extends from Moho to the core-mantle boundary at

2900 km depth. The whole of the mantle is now considered to be essentially
solid and to a large extent radially homogeneous. The compressional wave
velocity increases from about 8 km/s just beneath the Moho to 13.7 km/s at
the core-mantle boundary (Fig. 5). The mantle may be subdivided into the
upper mantle, including the non-crustal lithosphere and the asthenosphere,
and the lower mantle. The upper mantle extends to a depth of about 700
km, where the velocity gradient suddenly decreases, and contains several
10

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.

discontinuities. There is unquestionable seismological evidence of interfaces,

e.g., at depths of 400 and 650 km. They all are less precisely determined
than the Moho and, hence, some workers prefer to work with models
containing transition zones or layers of a thickness of the order of, say, 50
km rather than with definite or sharp discontinuities. It is assumed that
within the transition zones the velocity increases with depth more rapidly
than in the surrounding layers. It is worth mentioning that recent research
provides good evidence that the 650 km discontinuity is sharp, e.g. short-
period sharp reflections in P'dP' (for nomenclature see Section 4.2.2 or
Table 3). On the other hand, the 400 km discontinuity is not sharp. One of
the important features of the upper mantle is the world-wide existence of a
low-velocity layer (LVL) between about 100 and 250 km below the surface.
Within the LVL, the rocks are partially molten, the rigidity is low, the
attenuation is the largest of the whole mantle and seismic wave velocities fall
by about 6% when compared with the velocity just under the Moho. It is
obvious that the LVL plays an important role in the propagation of seismic
waves. The lower mantle extends from some 700 km depth to the
core-mantle boundary at 2900 km depth, first recognized by R.D. Oldham
in 1906 and accurately located by B. Gutenberg in 1913. Seismic velocities
 11

in the lower mantle increase gradually with increasing depth although at a

significantly lower rate than in the upper mantle. There are no distinct
reflectors/refractors in the lower mantle.

3.3 CORE

Below the core-mantle boundary is the core of the Earth with an

approximate radius of 3500 km. The boundary represents a sharp thin
discontinuity in physical properties such as a precipitous fall of the
compressional-wave velocity from 13.7 to 8.1 km/s and cessation of shear
waves (Fig. 5). In spite of great observational efforts, no shear waves that
have traveled through the core have yet been identified on seismograms. It is
generally accepted that shear waves cease to exist at this depth due to the
fluid character (no resistance to shear, i.e. no shear strength) of the core.
Seismic wave studies led to a subdivision of the core into an outer core,
which in relation to seismic waves acts as a liquid and an inner core which
acts as a solid. Some early workers claimed originally that the inner and
outer core were separated by a transition layer about 150 km thick within
which the compressional-wave velocity declines sharply. Recent studies do
not show this transition layer and advocate the existence of a rather sharp
discontinuity in the compressional-wave velocity at the bottom of the outer
core. The compressional-wave velocity in the inner core is significantly
higher than that in the surrounding outer 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

4.1 BASIC TYPES AND ESSENTIAL PROPERTIES

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.

As far as surface waves are concerned, we shall here limit ourselves to a

brief description of two basic types, namely to Love waves and Rayleigh
waves which are often the dominant wave types seen on actual seismograms.
It is usual to denote Love waves by LQ and Rayleigh waves by LR where L
stands for long (i.e. long waves), Q for Querwellen, an alternative name from
German for Love waves, and R for Rayleigh waves. LR and LQ waves
propagate along the Earth's free surface or in layers bounded by velocity
 15

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.

Amplitudes of LR and LQ waves are largest at or near the surface and

decrease rapidly (roughly exponentially) with depth. Consequently, shallow
(crustal) earthquakes generate large surface waves but with increasing focal
depth, surface waves become smaller and smaller. For crustal earthquakes,
surface waves usually dominate the seismogram while for deeper shocks (h >
100 km) they often become insignificant. Obviously, this fact provides the
analyst with a powerful tool for a quick (at first glance) and reliable
discrimination of shallow earthquakes against deep shocks.

Both Love and Rayleigh waves exhibit an important property called

velocity dispersion or, for short, dispersion. The velocity of propagation of
dispersed surface waves is not period (or frequency) invariant, as is the case
of Ρ and S waves, but increases with increasing wave period (normal
dispersion). In practice, this means that the long surface waves approach the
station first and are recorded ahead of the "slower" shorter waves. Hence,
ideally the seismogram of surface waves LQ or LR will start with rather
long-period motion which gradually, as time increases, will turn into shorter
and shorter periods.

In Rayleigh waves, the motion of rock particles follows a retrograde

elliptical orbit in a vertical plane pointed in the direction of the generating
earthquake (see Fig. 6). In Love waves, there is no particle motion in the
vertical direction. Particles move in a horizontal plane at right angles to the
direction of the wave propagation (see Fig. 6). As in the case of body waves,
different polarization of particle motions is an important clue which often
makes it possible to distinguish between different surface waves LQ and LR.
For example, as follows from the above description, vertical-component
seismographs cannot record Love waves. Another important clue is that LQ
waves travel somewhat faster and therefore precede LR waves on
seismograms.

If the studied earthquake exceeds a certain magnitude, its seismic waves

can be recorded by sensitive seismographs placed all around the world on the
surface of the Earth, in boreholes, abandoned mines, on the ocean bottom,
16

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.

Since various types of seismic waves propagate with different velocities,

they arrive at the recording site well separated in time so there should
theoretically be no difficulties in their identification on seismograms. This
behaviour has also been demonstrated by making use of the Greek earthquake
record shown in Fig. 7. However, it has to be emphasized that in this
 17

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.

The ground is practically always in motion. Various human activities such

as traffic, construction work, industries etc generate so called cultural noise
with dominant frequencies usually above 1 Hz. A similar type of noise is
also generated by action of wind, smaller water basins or rivers e t c Various
interactions between atmospheric effects, oceans and the solid Earth give rise
to microseismic noise sometimes also called ocean microseisms. Dominant
frequencies of microseisms occupy a broad low-frequency range from less
than 0.01 Hz to, say, 0.5 Hz, i.e. periods from 2 s to more than 100 s. The
most common microseisms have more or less regular periods of about 6 s.
18

Cultural noise is recorded with standard instruments at epicentral distances

usually not exceeding several tens of kilometers and is, therefore, only of
local importance. Microseisms, on the other hand, can travel many hundreds
of kilometers and hence, are a continental phenomenon. They are correlated
with stormy weather conditions in some adjacent oceanic regions and may
persist from several hours to several days or weeks. Oceanic microseisms
often show strong seasonal variations in both the amplitude level and the
dominant period.

A variety of sources radiate a variety of noise types and to list commonly

valid characteristics is rather difficult. Generally speaking, cultural noise due
to its relatively high frequency content, affects records of near events
(epicentral distance less than about 1000 km). Microseisms, on the other
hand, interfere with records of distant events and make the interpretation
difficult and sometimes even impossible. Examples of "noisy" seismograms
are shown in Fig. 8.

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).

To complicate matters further, seismic waves encountering a discontinuity

are reflected and/or refracted (at the Earth's free surface, seismic waves are
reflected downward) and an incident Ρ or S wave gives rise to both Ρ and S
 19

waves (mode conversion). Thus, a seismogram from a distant earthquake will

often show a number of more or less distinct waves, commonly called
phases, distributed in time, which have traveled along different propagation
paths and which were subjected to different mode conversions Ρ to S or S to
P.

Wave energy is also scattered by velocity heterogeneities crossing the

propagation path. For the given wave period, the scatter affects S waves
more than Ρ waves. Scattered waves reach the Earth's surface after the Ρ
wave and contribute to the buildup of Ρ coda which in its turn obscures the
later phases. Hence, except for the first Ρ onset, all later arrivals are
contaminated by codas of preceding phases so that on the record there is
virtually no interval of quiescence between individual arriving phases.

Before we proceed further, it is worthwhile to summarize the most

important characteristics of seismic waves which deserve our attention, and
which are invaluable in any seismogram interpretation. Firstly, different
waves travel with different velocities. At any epicentral distance, Ρ is
recorded first, followed by S, LQ and LR. Secondly, different waves are
polarized in a different way (P linearly, LR elliptically, etc). This provides a
means of identifying phase types. Thirdly, various phases show certain
characteristic features (amplitude, period, dispersion, etc) which again are of
primary importance for correct interpretation.

4.2 PROPAGATION PATHS

The propagation of seismic waves through the Earth's interior is governed

by exact mathematical laws similar to the laws of light waves in optics. If
the propagation velocities and other elastic properties were uniform
throughout the Earth, seismic waves would radiate from the focus of the
earthquake in all directions through the Earth along rectilinear paths or rays.
In general, however, the wave velocity increases with depth and
consequently, seismic rays are not straight lines but lines curved with the
concave side upward providing the shortest time-path through the Earth. To
be able to simulate various discontinuities in the Earth and at the same time
to simplify the associated ray geometry, we shall, hereafter, assume the Earth
to be a sphere, made of a finite number of concentric spheroidal
homogeneous shells. Elastic properties vary from shell to shell but remain
constant within each shell. For a spherical Earth model, it is common
practice to express the distance between the focus and the recorded station as
the angle (denoted Δ) subtended at the center of the Earth by the arc
between the source and receiver (1° = 111 km).

In the following discussion, we introduce, for the sake of clarity, three

20

different categories of seismic events. The classification is based upon the

distance between the event and the recording site (i.e. upon the epicentral
distance) which in its turn governs propagation paths along which seismic
waves travel through the Earth's interior. The main reason for this
classification, which does not provide any sharp line of demarcation, is that
seismic waves from different categories may be discriminated from each
other due to their different appearance on seismograms.

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.

4.2.1 Crustal waves; recording distances 0-10°

In order to explain the structure of seismic records made at epicentral

distances between 0 and 10° (some workers use the limit 1000 km), let us
first assume a much simplified structural model of the crust, depicted in Fig.
9. Note that for the distances considered here, we can neglect the effects of
curvature of the Earth's surface.

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

called the critical angle, is denoted i . Corresponding Ρ and S waves called

c

head waves, recorded at S are labeled Pn and Sn, respectively. Waves

3

propagating along discontinuities separating two layers with two different

velocities move with the higher of the two velocities. Thus, Pn and Sn waves
depicted in Fig. 9 travel with velocities of the uppermost mantle. As can be
seen in the figure, Pg and Sg exist for all epicentral distances from Δ = 0
and outwards whereas Pn and Sn phases cannot be observed at distances
shorter than that corresponding to the location of the station S (Fig. 9), i.e. 2

at distances shorter than the critical distance, which for the continental crust
is about 100 km.

Structural models, like the model depicted in Fig. 9, should be as simple

as possible to make the seismogram interpretation manageable. On the other
hand, the model should also be accurate enough to reflect the actual
22

structure. Evidently, these are two opposing constraints and a proper

compromise must, in each particular case, be made to make the interpretation
possible. It should be stressed that in the above presentation and in Fig. 9, a
number of simplifications has been made.

Firstly, a homogeneous, one layer, crust will in many cases be a rather

poor approximation of the true structure. It is common to employ two
crustal layers, separated by the Conrad discontinuity, to interpret crustal
phases. In special studies, multilayered crustal models are used.

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.

Lastly, the assumption of Pg or Sg traveling as direct waves (P or S) over

large distances, as shown in Fig. 9,_is again a gross simplification. The true
case is that direct waves, denoted Ρ and 5, are recorded only from local
events, i.e. at very short epicentral distances usually not exceeding several
tens of kilometers. Pg and Sg waves are then understood as channel or
refracted waves traveling along less pronounced boundaries within the granitic
layer.

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.

layer in the oceanic crust. Similarly, earthquakes originating in the lower

crust, beneath the Conrad discontinuity, do not produce Pg or Sg phases.
Hence, first arrivals on records from these earthquakes will be Pn or P*.

For ease of phase identification, rather than relying on the epicentral

distance, which of course may not be available, we can make use of time
differences between arrivals of various phases (see Chapter 5). For example,
if the arrival-time differences S-P is less than about 20 s, the first wave
within the Ρ and S group to arrive at the recording site is probably Pg (or
P) and Sg (or S), respectively. If on the other hand, the difference is more
than 25 s, the first arrival is most likely Pn. Details obviously depend upon
the true structure and upon the focal depth. It should be emphasized that
only seldom are all the above phases identified on one record. The usual
case is that one or several of these waves are too weak or hidden in the
background noise to be discernible on the seismogram.

With a certain time delay following the Ρ phases, proportional to the

epicental distance, the crustal S waves arrive in the same order as Ρ waves.
Thus, for local events (earthquakes, mine explosions, quarry blasts, etc) the
order of S onsets will be Sg, S*, Sn while for events from distances larger
than about 300 km we observe first Sn followed by 5* and Sg.

As far as Ρ or 5 waves reflected at Moho are concerned, corresponding

onsets on seismograms are rather scarce and difficult to identify. The best
chance to record PmP or SmS is at very short distances where the
24

contamination by Pg and Pn (or Sg and Sn) is not severe. In rare instances,

additional so called depth phases, arriving between Pn and Pg, are present on
seismograms. These leave the focus as Ρ waves, travel upward with a small
angle of incidence, are reflected as Ρ at the free surface and continue further
as Pn. Notation for this phase is pPn. Similarly, sPn denotes a depth phase
leaving the focus as 5, converted through the reflection at the free surface
and continuing as Pn. Ray paths of pPn and sPn are depicted in Fig. 11.
Both pPn and sPn are of great importance in focal depth estimations but at
the same time rather difficult to identify on actual records.

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.

Short-period S waves multiply reflected between the free surface and

Moho, or between other crustal velocity discontinuities, interfere with each
other and give rise to a wave group labeled Lg which follows the Sg arrival.
The subscript g again refers to granitic layer. At distances of several hundred
kilometers and larger (continental paths), Lg waves, which propagate as
guided waves, supercritically incident on the Moho and multiply reflected
within the crust, and with a typical velocity of about 3.5 km/s, may dominate
the seismograms, especially the horizontal channels. Lg is usually recorded at
epicentral distances of about 5° and larger. Cases are known where Lg
propagated over distances of several thousand kilometers (e.g. from
USSR-China border region to Sweden, see Plate 33) and were recorded as
prominent phases on the seismograms.

Near-surface regional events (earthquakes, industrial explosions, rockbursts

etc) also generate short-period surface waves of Rayleigh type, labelled Rg.
The presence of short-period Rg in the seismogram is a reliable indicator of
a very shallow event with focal depth of the order of one or a few
 25

kilometers. On the other hand, if short-period Rg waves are absent (near

station, epicentral distance of several hundred kilometers or less), we are
concerned with a deeper natural event, i.e. with a crustal earthquake at a
depth most likely between about 5 to 25 km, since all types of man-made
events as well as triggered mine tremors can be excluded. Short-period Rg
waves, which travel as guided waves through the crust across continental
paths with velocity of 3 km/s or slightly higher, are exposed to more
effective attenuation when compared with crustal body waves and their range
of propagation is therefore limited to distances less than about 600 km
(B&th, 1983). However, at short epicentral distances, of say, less than 100 or
200 km, the Rg phase from a near-surface event often dominates the recorded
wave train (see Plate 5).

Seismic waves from local and regional earthquakes of low or moderate

magnitude are of short period and therefore almost exclusively recorded by
short-period seismographs. The seismogram length depends upon the
magnitude but generally does not exceed 5 minutes or so. The number of
clear pulses seen on the record, indicating arrivals of various Ρ and S waves,
is often higher than one would expect from models displayed in Figs, 9, 10,
and 11, demonstrating the departure of models used from the true structure.
For continental travel paths, the most prominent phase is ususally Sg, best
recorded by horizontal-component instruments. Generally speaking, Sg arrivals
start with large amplitudes which successively decrease as the time increases
forming the coda of the event. Coda duration is related to the magnitude. Rg
phases, best recorded on vertical-component seismograms, often display a
clear dispersion. As an example, a record from a regional earthquake is
displayed in Fig. 12. The first discernible phase is Pn, weakly recorded on
the vertical component. It is followed by Pg, Sn and Sg which are
recognizable on all three channels. Largest amplitudes are exhibited by the
Lg wave trains on the two horizontal channels. The focal depth of about 15
km prevents the development of Rg waves. Note the high-frequency character
of all recorded phases.

Strong events (magnitudes about 6 and larger) recorded at local or regional

distances (Δ < 10°) will produce seismograms with duration of several hours.
Associated large amplitudes of ground vibrations often saturate the recording
system (clipped records) and the resulting seismogram is not of much use
except for measuring the arrival time and polarity of the very first recorded
phase. In this respect, digital systems with higher dynamic range are superior
to analog instruments (for more details see Section 6.4).

Similar to the LVL in the upper mantle, there is also a low-velocity

channel in the deep ocean. Depending upon the salinity and temperature of
water, the sound velocity decreases from the sea surface to a minimum of
about 1.5 km/s at about 700 - 1300 m depth and increases again from that
26

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.

Island and coastal seismographic stations frequendy record these waves

termed Τ waves (tertiary waves) arriving after Ρ and 5 and characterized by
propagation within the oceans as ordinary sound waves. Seismic waves
emitted by earthquakes near the sea bottom or by submarine volcanic
eruptions are refracted through the sea floor and propagate as sound
(longitudinal) waves through the ocean. The propagation of Τ waves,
generally by the SOFAR channel (Bullen and Bolt, 1985) or by multiple
reflections between the sea floor and the sea surface (B&th and Shahidi,
1974), is very efficient and observations at distances as large as about 80°
have been reported (see Plate 49). First observation of Τ waves was made by
D. Linehan in 1940.

Τ waves are best recorded by ocean-bottom seismometers (OBS) and by

coastal and island stations (see Plate 14). However, instruments deployed
further on land sometimes also record clear Τ waves after a water-land
 27

conversion of sound waves into P, S or surface waves propagating over the

land portion of the total transmission path. If this is the case, the labeling is
TPg, TSg and TRg, reflecting the fact that the path of propagation over the
land is within the crust. An example of recorded TSg phase is given in Fig.
13. Τ waves are short-period waves, with periods usually less than 1 s,
recorded exclusively by short-period seismographs. On records, they often
exhibit rather monochromatic oscillations with a gradual increase and
decrease of amplitudes of total duration up to several minutes (Fig. 13).
Some workers (B&th and Shahidi, 1974) report inverse dispersion observed in
Τ wave trains. When compared e.g. with Ρ waves, there is no sharp onset in
the Τ wave group which obviously creates difficulties when reading the Τ
arrival times or when identifying phases within the Τ wave group. In general,
there is great variety in the appearance of Τ phases due to the dependence
upon the bottom topography in the vicinity of generation, oceanic
stratification and water-land conversion and transmission (Bath and Shahidi,
1974).

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

recorded Τ phase amplitudes often exceed those of associated Ρ wave by a

factor of up to 30.

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.

It is not always possible for analysts to identify correctly crustal phases

from the records of a single station, although this may be easier if several
stations of a network are read together. If there is doubt about correct
interpretation, a phase should simply be identified as Ρ or 5.

4.2.2 Body waves; recording distances 10-103°

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

mantle and corresponding seismograms are relatively simple.

To explain various features of waves traveling through the mantle, let us

consider the Earth's cross section, a surface focus event and travel paths of
the more important body waves depicted in Fig. 14. Note that for the
distance range of 10-103° studied here, we have to introduce the spherical
shape of the Earth. A flat Earth model, used in the case of crustal waves, is
no longer appropriate. As mentioned above, seismic rays in the true Earth are
not straight but bent upwards due to the velocity increase with depth. This
means that waves traveling to more distant stations penetrate the earth to
greater depth than those traveling to near stations. Because of greater depth,
the velocity of propagation is larger, i.e. Ρ and S waves reach the more
distant stations more quickly then might be expected. In other words, there is
a non-linear relation between the distance and travel time for Ρ as well as
for 5 waves (see Chapter 5).

At distances around 10°, Ρ η and Sn become difficult to identify in the

records, except in some shield areas and other regions with relatively uniform
structures. Instead, teleseismic Ρ and S phases become visible on
seismograms. Ρ is usually stronger on the vertical component, while S is
more clearly seen on horizontal components. S often exhibits wave trains
with longer periods when compared with corresponding P. Large-amplitude S
waves are often observed at distances of up to about 100°.

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

free-surface reflected phases become very distinct. At distances around 100°

and larger, PP and SS often belong to the largest recorded body waves.

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.

A symbol c is used to indicate a single upward reflection, i.e. a reflection

back into the mantle from the outer core-mantle boundary. For instance, ScP
(Fig. 14) corresponds to an 5 wave which travels down from the focus,
strikes the boundary, is reflected and converted into the Ρ type wave and
finally is recorded at the Earth's surface as ScP. A straightforward extension
provides PcP, ScS and PcS. Because these phases emerge steeply, ScP is
usually stronger on vertical components than PcS. Large reflected core phases
are usually recorded at shorter epicentral distances, say at 40° or less. At
distances around 39°, ScP and PcS (surface foci) are often contaminated with
the arrival of direct 5 and the phase separation is difficult. When the ray
path of PcP grazes the outer core boundary, the combination of direct Ρ and
PcP is called Ρ diffracted. This case is discussed in more detail below. Core
reflected waves together with PmKP (see Section 4.2.3 for notation), recorded
from earthquakes at a wide range of distances and focal depths are used to
study the properties of the core-mantle boundary.

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.

4.2.3 Body waves; recording distances 103° and larger

It was noticed in the early days of observational seismology that amplitudes

of direct Ρ waves decay dramatically at distances larger than 100°. The
short-period Ρ waves reappear consistendy on records first at distances of
about 140° and larger. Correspondingly, the distance range 103°<Δ<140° is
called the shadow zone. Within this zone, there is no penetration of direct Ρ
waves due to the wave diffraction around the Earth's core (Fig. 16). The last
direct Ρ wave reaches the Earth's surface at an epicentral distance of about
103° where the shadow zone produced by the Earth's mantle commences. Ρ
waves traveling beyond this distance creep around (are diffracted) the
core-mantle boundary and lose a large part of their energy there, so that only
weak, diffracted Ρ phases are observed in this distance range. Similar to
dispersion (see Section 4.1), diffraction also depends on the wave period (or
frequency). The longer waves are diffracted more into the shadow zone than
shorter waves. The diffracted Ρ waves are labeled Pc (or Pdif, Pdiff).
Seismological centers like the National Earthquake Information Center (NEIC)
in Golden, Colorado, or ISC in England use Pdif. On seismograms, Pc waves
usually show small amplitudes, emergent or gradual onsets and the energy
shifts to longer periods. Long-period Pc are sometimes observed out to
distances of 160° or more. S waves are affected at the core-mantle boundary
in a similar way; the symbol Sc (or Sdif, Sdiff) is used for diffracted S
waves. The shadow zone for S waves on the side of the Earth opposite the
earthquake, extends over all epicentral distances from about 103° to -103°
(257°).

Direct Ρ and S waves and corresponding reflections are easily

distinguishable from the recorded surface waves. The former usually occupy
the period interval from, say, 1 to 5 s, while the latter show large amplitudes
(surface or shallow shocks) and periods in the interval from about 10 to 100
s. The period of 5 waves increases with distance and, in cases of multiple
reflection may reach even several tens of seconds.
As mentioned above, the limiting ray path which is tangential to the core in
the real Earth is that corresponding to an epicentral distance of about 103°. Ρ
and S waves recorded at distances greater than 103° graze or strike the
surface of the core, the wave pattern becomes rather complicated and
amplitudes decay dramatically. Seismic waves which leave the source with a
steeper descent than the grazing waves, strike the core-mantle boundary and
are subsequently divided into reflected and refracted waves. The former are
 33

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).

A Ρ wave penetrating the outer core is denoted by Κ (from German

Kernwellen for core waves). After traveling through the outer core and
following another partitioning (reflection or refraction) at the outer core
boundary it emerges at the Earth's surface. Thus, we can form the four
symbols for waves which have traveled through the outer core: PKP, PKS,
SKS and SKP. For example, the symbol PKS corresponds to a wave that
starts in the mantle as a Ρ wave, is refracted into the outer core as a Ρ wave
and is finally, after mode conversion, refracted back in the mantle as 5. The
phase SKP is stronger on vertical components than PKS. These phases have a
caustic near 130° and at this distance are often the only phase recorded on
short-period instruments. Some of the seismic rays traversing the core are
illustrated in Fig. 17. It has to be emphasized that the symbol Κ always
represents a Ρ wave since S waves do not enter the outer core. The notation
PKP is sometimes abbreviated as F ' .

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.

If the studied event is weak, then usually no Pc is observed in the entire

distance interval Δ > 103° and the first arrival seen on the record will be that
of PKP. At epicentral distances 105-120°, PKIKP usually provides the first
onset discernible on the seismogram.

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

adopted from travel-time charts with branches denoted AB, BC and DF

(Jeffreys and Bullen, 1967). Arrivals associated with these branches are then
labeled PKP^, PKP BC and PKP DF and correspond to PKP2, PKP1 and
PKIKP arrivals, respectively. The CD branch (PKP ) is related to PKiKP
CD

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'.

In the late 1960's, first observations of forerunners to P'P' were made.

These were interpreted as P'dP'. Analogous to PdP, P'dP' waves are not
reflected at the opposite surface of the Earth (as is the case of P'P') but at
some discontinuity in the upper mantle. For example, P'650P' travels from
the hypocentre to the other side of the Earth, where it is reflected back to
the station from a layer 650 km below the surface. P'650P' passes through
the core twice and on the seismogram precedes P'P' by about 2 minutes.
Some close precursors within about 30 s of the main P'P' arrival can arise
from asymetrical reflections.

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

4.2.4 Body waves from intermediate-focus and deep-focus earthquakes

In the early days of observational seismology, the focal depth of recorded

earthquakes was often a topic of speculation even though there were strong
indications (such as surface faulting or the limited area of destruction) that
many of the earthquakes must have been rather shallow events. On the other
hand, early in this century several investigators (Pilgrim, 1913; Turner, 1922)
found a number of events with focal depths greater than 100 km. Somewhat
later, Japanese scientists (Wadati, 1927; Shida, 1937) presented evidence that
Japanese earthquakes occur at practically all depths down to 500 km. They
based their conclusions upon observed S-P arrival time differences, intensity
distributions and different appearances of intermediate- or deep-focus
earthquakes and of those that take place at shallow depth. Wadati noted very
early that the seismograms of intermediate-focus and deep shocks display
rather impulsive and large S phases, shorter predominant periods and less
well developed codas. In this context, it is perhaps interesting to mention that
during the 1920's the Japan Meteorological Agency (JMA) already operated
an excellent regional network of seismographic stations (Frohlich, 1987).

Later studies confirmed conclusions from Wadati's pioneering work that

intermediate and deep-focus shocks produce simpler seismograms with
exceptionally well recorded impulsive body waves while surface wave
amplitudes decrease as the earthquake becomes deeper. Strong depth phases,
such as e.g. pP and sS, are also frequently very distinct on records from
deep events. However, the duplication of principal phases by surface
reflections often complicates the seismogram interpretation. Another important
characteristic that accentuates the difference between shallow and deep shocks
is the pattern of aftershocks. While large shallow earthquakes are usually
followed by numerous aftershocks, deep events (which may be multiple
shocks) virtually never show well developed aftershock series.

As an example, Fig. 20 displays several records made at teleseismic

distances from the Sea of Okhotsk earthquake which occurred at a depth of
580 km. Note the rather impulsive appearance of F , PcP, pP and especially S
and ScS (E-W component, short-period trace) which are all easily indentified
on the seismogram. Practically no surface waves were recorded from this
event.

Strong intermediate-focus and deep earthquakes occur in several different

seismically active regions. Among these are: island arcs such as
Tonga-Kermadec Islands, the Marianas, New Hebrides Islands or the Aegean
arc; continental margins with deep ocean trenches like Central America and
western South America; mountain chains e.g. Himalayas (Hindu Kush) or
Carpathians. About one fifth of all reported earthquakes take place at a focal
depth exceeding 70 km. Among the deepest known earthquakes are three in
 39

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

deepest two, cannot be regarded as well established. Occasionally, deep-focus

earthquakes occur in rather unexpected geographical areas. For instance, up to
1954, it was generally accepted that the geographical extent of deep quakes
is limited to the Circum-Pacific belt. However, on March 29, 1954, an
isolated major earthquake occurred in southern Spain at a depth of 630 km.
A smaller earthquake occurred at the same focus on January 30, 1973.

Even though several alternative hypotheses to explain deep earthquakes have

been launched in the past (e.g. contraction of the Earth due to cooling),
today, nearly all seismologists agree that deep and intermediate-focus events
are associated with the subducting lithosphere which fits nicely with the idea
of plate tectonics. It is of course, still possible that individual shocks, e.g.
that of March 29, 1954, may not be related to subduction. Notwithstanding
the unifying frame of plate tectonics, it is likely that deep and shallow events
are generated by fundamentally different modes of rock failures. For example,
it is possible that deep shocks are not associated with dislocations along
quasiplanar fault surfaces. Some researchers now ascribe deep quakes to
phase transitions taking place in the upper mantle or in other words to
sudden voluminous changes (densification) due to the collapse of olivine (a
major mineral in the Earth's mantle) molecules from the low-pressure form
into a more dense form. However, the relation between phase transitions,
generation of deep earthquakes and the style of mechanical failure associated
with these quakes is still poorly understood and debated. Other plausible
models explaining the mechanism of earthquakes at great depth, such as
dehydration of minerals or an abrupt build up of the frictional heat, have also
been proposed (see e.g. Frohlich, 1989).

On the whole, the most distinctive features of a large intermediate-focus and

deep earthquake recorded at teleseismic distances pertinent to seismogram
interpretation are the seismogram simplicity, small amplitudes or even
absence of surface waves and an impulsive shape of body waves. With some
elementary experience, all these characteristics may often be revealed as the
eye scans the seismogram. In a more retrospective-type interpretation, the
absence of aftershocks will support the classification of the shock as a deep
or intermediate-focus event.

4.2.5 Surface waves

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

that interaction of body waves along velocity dicontinuities generates various

types of surface waves and presented some of their essential properties. In
the following lines we complement the previous discussion with more details
pertinent to interpretation of recorded surface waves.

In 1885, Lord Rayleigh proved theoretically that a special type of surface

wave (now called Rayleigh wave and labeled LR) can propagate along, i.e.
beneath, the Earth's free surface. Contrary to other types of surface waves,
propagation of LR waves is not limited to layered media; they can also be
transmitted through a homogeneous half-space (semi-infinite medium). As
discussed earlier (see Section 4.1), the particle motion of Rayleigh waves
follows a retrograde elliptical orbit in the vertical plane containing the
direction of propagation. At the Earth's surface, the amplitudes in the
vertical and horizontal directions are related roughly as 3:2. Hence, Rayleigh
waves are usually best seen on vertical-component seismograms. Amplitudes
of LR waves decrease rapidly (exponentially) with increasing depth. For
example, at a depth equal to one wavelength, the vertical and horizontal
amplitudes fall to 0.11 and 0.19 of their free-surface values, respectively. The
velocity of Rayleigh waves in a homogeneous medium, c , lies between 0.87
R

and 0.96 times the 5-wave velocity, v , of the half-space. For many rock
s

materials the Poisson ratio is approximately 0.25 which leads to a relation

c = 0.92v .
R 5

In 1911, A.E.H. Love solved the theoretical problem of wave propagation

through a thin superficial layer superimposed on a homogeneous half-space.
Assuming that the 5-wave velocity in the layer is lower than that in the
material below, another type of surface wave (called Love wave and
designated LQ) can be transmitted through the layer without any significant
penetration of energy into the lower medium. Love waves may, therefore,
also be considered as channel waves, in this particular case transverse waves,
trapped in the superficial layer. In contrast to Rayleigh waves, Love waves
show no vertical motion since particles excited by propagating LQ waves are
polarized in the horizontal plane perpendicular to the direction of propagation.
Consequently, traces of LQ waves should be looked for on the
horizontal-component seismograms. Love waves propagate faster than
Rayleigh waves with velocities limited by 5-wave velocities in the layer and
the half-space. The mean velocity of propagation is 4.43 and 3.97 km/s for
LQ and LR, respectively (Bullen and Bolt, 1985).

The analyst has essentially two clues to distinguish LQ waves from LR

waves. Firstly, the different particle motion. While Rayleigh waves usually
show the largest amplitudes on vertical-component records, Love waves are
best displayed on horizontal-component seismograms. Secondly, the different
propagation velocities. Both LQ and LR propagate slower than Ρ or 5, but
since LQ propagates faster than LR, they are recorded ahead of LR waves.
42

Surface waves traveling through layered media often show appreciable

normal dispersion which as time goes on, continually changes the shape of
both LQ and LR. Due to the dispersion, the original appearance of the wave
train becomes disturbed on the seismogram by long period waves advancing
toward the beginning of the wave train as it travels through the medium.
This rather special behaviour, leads to the concept of phase velocity, c, and
group velocity, U. The former is the velocity with which a wave of single
frequency (monochromatic signal) propagates. The latter is the velocity of
travel of the wave train energy, i.e. the velocity of the wave train envelope.
Plots of wave velocity as a function of period (frequency) are called
dispersion curves. Figure 21 displays empirical average group velocity
dispersion curves for fundamental mode Love and Rayleigh waves. Observe
that the shape of curves in Fig. 21 depends strongly upon the character of
the propagation path. In the period range from about 20 to 60 s, continental
travel paths exhibit dispersion curves for LR waves with group velocity
gradually increasing from 2 to almost 4 km/s. For LQ waves, the group
velocity starts at about 2.5 km/s and reaches values close to 4.5 km/s at
periods of several hundreds of seconds. As follows from the figure,
dispersion curves associated with oceanic travel paths show a rather abrupt
change in the short-period range. At periods around 15 s, the velocity of LR
increases sharply from about 1.5 to more than 3 km/s. For LQ wave, a
sudden velocity rise from about 3 to 4 km/s is seen at periods around 7 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

An experienced analyst will distinguish between recorded surface waves that

have traveled along pure oceanic or continental paths. Dispersion
characteristics of oceanic routes give rise to long wave trains with rather
slow and sometimes hardly visible period change over relatively long (5-10
minutes) record segments. Several typical seismograms are displayed in Plates
34 and 50. In contrast, continental paths generate a characteristic fast period
decrease with time, which is often easily recognized by inspecting several
minutes of the records (see e.g. Fig. 7 and Plates 22 and 39). As indicated
above, the exact shape of dispersion curves depends upon the traversed
structure. This means that available empirical dispersion curves for LQ or LR
waves provide the researcher with remarkably effective probes for studying
the structure of the traversed medium (so called inverse problem).

As can be seen in Fig. 21, dispersion curves show a rather complicated

pattern with several local minima and maxima. Surface waves traveling with
these minima or maxima group velocities are called Airy phases. On
seismograms, an Airy phase is characterized by a constant-frequency compact
wave train, often with a remarkable amplitude buildup of dispersed surface
waves traveling by fundamental-mode propagation.

At short epicentral distances, it is difficult to identify LQ and LR waves

because they are often contaminated by large-amplitude 5 waves. On the
other hand, at large distances, the identification is rather simple since LQ and
LR waves dominate the record (shallow events) and are significantly delayed
with respect to S waves. On seismograms, surface waves may be spread over
several hours.

Figure 21 displays another interesting feature which should be dicussed in

more detail. Observe that the dispersion curves for LR waves (both the
oceanic and continental paths) show a local minimum for periods around 200
s. In the period range from approximately 50 to 200 s, the group velocity is
monotonically decreasing with increasing period. Physically this means that in
this period range, long-period Rayleigh waves follow the laws of inverse
dispersion. Observations of this interesting phenomenon are rather scarce,
however one example is displayed in Plate 30.

Periods of the largest (maximum amplitude) recorded surface waves show a

clear positive correlation with epicentral distance. For example, for distances
10, 50 and 100°, the expected minimum periods, 7V., of the largest
continental Rayleigh waves are of the order of 7, 13 and 16 s, respectively
(Willmore, 1979). For oceanic passages, the periods can be somewhat longer.
Also the time occurrence of the beginning of the maximum movement, T Rmax

(Rayleigh wave), with respect to the first onset of Ρ waves, T , is obviously

P

distance dependent. Table 1 gives the time differences T - T for various

RmaK P

epicentral distances Δ (after M. BSth, 1947, abbreviated). Hence, the position

44

of the maximum amplitude in the LR wave train, with respect to the Ρ

arrival, and its period offer the interpreter important information on the
epicentral distance of the earthquake.

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

10° 6.2 min 7 s 60° 35.0 min 14 s 110° 63.7 min

15 9.1 8 65 37.8 115 66.6
20 12.0 9 70 40.7 120 69.5
25 14.8 75 43.6 125 72.3
30 17.7 10 80 46.5 16 130 75.2
35 20.6 85 49.3 135 78.1
40 23.5 12 90 52.2 140 81.0 18 s
45 26.3 95 55.1 145 83.8
50 29.2 100 58.0 150 86.7
55 32.1 105 60.8 160 92.5

As mentioned earlier, surface-wave amplitudes are large only close to the

Earth's free surface and decrease, broadly speaking, rather rapidly with
increasing depth. Consequently, a shallow-focus earthquake will usually
generate large dominating surface waves, often exceeding the amplitudes of
recorded body waves, while a deep-focus earthquake of the same magnitude
will generate abnormally small (insignificant) surface waves. This feature
obviously provides the interpreter with a viable tool to discriminate, at first
glance, between shallow events and deep-focus earthquakes.

With the advent of long-period instruments, comparatively long-period

surface waves, so called mantle waves, have been observed from large
distant shocks. These waves can be of either Love- or Rayleigh-wave type
with approximate periods of somewhat less than one minute to several
minutes. The speed of the LQ type mantle waves is often nearly constant at
4.4 km/s (see Fig. 21) and the wave has an impulsive shape on the
seismogram. The LR type mantle waves travel with a velocity between 3.6 an
4.1 km/s. Since the wavelenghth of mantle waves varies from several
hundreds to more than thousand kilometers, a large part of the Earth's mantle
is affected by these waves. An interesting feature of mantle waves is their
repeated appearance on records, which is due to their multiple travel around
the Earth. The LQ type mantle wave was given the label G (after B.
Gutenberg) and- the LR type mantle wave, the label R. The older
nomenclature sometimes uses W instead for R (from German
Wiederkehrwellen, meaning repeated waves). G waves that propagate the
 45

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

The difference, therefore, becomes indicative of the epicentral distance. The

time that elapses between the arrivals of Ρ and 5 is not the only possibility
to determine the epicentral distance; combinations such as PP-P, among
others, are sometimes used as well. The travel-time difference, e.g. S-P, is
measured from the seismogram and the corresponding epicentral distance is
uniquely determined from the travel-time chart (see Fig. 22) or from the
corresponding tables. Assume that we interpret a recording from a
shallow-focus earthquake made at a seismographic station A and that we
succeeded to identify on the seismogram the arrival of Ρ at 16 18 16.0 and
the arrival of 5 at 16 24 45.3. The travel-time difference, S-P, of 6 min 29.3
s can then be vertically fitted into just the right interval between the Ρ and 5
curves in Fig. 22. By reading down onto the epicentral distance scale, one
can ascertain that the earthquake was 44° away from station A. Combining
three or more stations, the epicentre is located by making use of the
48

swinging arcs method. It is then a matter of brief algebra to calculate also

the origin time of the earthquake. Obviously, with many stations and
required high precision, the source location is usually determined by
computer, and in practice, locations are carried out by comparing absolute
arrival times of phases at different stations, rather than using S-P intervals.

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.

Time of travel of a certain body-wave type and the curvature of the

travel-time curve depend on the velocity distribution along the propagation
path through the Earth's interior. Note, for example, that the observed
curvature can only be explained by structural models where the density and
in turn the velocity increase with increasing depth. Hence, besides its value
in locating earthquakes and in determining their time of occurrence,
travel-time tables also provide the seismologist with important information
concerning the structure of the Earth's interior. In this context, it is worth
mentioning that in contrast to body waves, travel times for surface LQ and
LR waves are represented by straight lines indicating that these waves
propagate along some (surface) layers with constant velocity.

Travel-time tables provide the key for successful phase identification. In

interpretations made by professional analysts not only all clear and significant
 49

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

The situation is quite different when analyzing seismograms from nearby

earthquakes, i.e. shocks at epicentral distance of 10° or less. In contrast to the
mantle and core, crust shows significant regional variations in structure which
heavily influences travel times of all crustal waves. For example, as
mentioned in Chapter 3, in continental areas, the crust is about 40 km thick
and the Ρ velocity varies from about 6 to 7.5 km/s. On the other hand, over
some oceanic areas, the thickness is only 5 km, the layering practically
disappears and the Ρ wave speed is found to be around 8 km/s. Thus, to be
able to identify the arriving waves on seismograms and to determine correctly
the source parameters (location, origin time), accurate regional tables,
applicable in the specific region must be on hand. Further allowance is
sometimes made for local geological structure beneath the station, the height
of the station above sea level, etc.

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).

In 1968, a new set of "world-wide" or "world-average" tables was published

by E. Herrin and co-workers (Herrin et al., 1968). While JB tables are based
merely on earthquake data, Herrin and his colleagues employed both
earthquake and large underground explosion data. Controlled locations and
origin times of the explosions, improved seismological instrumentation and
access to large computers were the factors which led to expectations of
significantly increased precision over the standard JB tables. However, the
undertaking demonstrated that the JB tables are in need of only minor
corrections, to be more specific, of travel-time reductions by approximately
 51

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

35 6 56·ι 449 6 88 6 32-9 27'2 6

36 7 θ4·6
85
5ΐ·3
59-8
85
53-4
85
6
3

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

49 8 5θ·3 8 454 8 8·7 8 32-2 8 2 8 8 ΐ4·ι

6
5
77 77 3
77 7 75

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

37.5 7 14.73 7 12.51 7 9.34 7 8.36 7 5.93 7 3.52 7 1.13

8.46 8.45 8.44 8.44 8.42 8.41 8.40
38.0 7 18.96 7 16.74 7 13.56 7 12.58 7 10.15 7 7.73 7 5.33
8.42 8.42 8.41 8.40 8.39 8.38 8.37
38.5 7 23.17 7 20.95 7 17.76 7 16.78 7 14.34 7 11.92 7 9.51
8.39 8.38 8.37 8.37 8.36 8.34 8.33
39.0 7 27.37 7 25.14 7 21.95 7 20.97 7 18.52 7 16.09 7 13.68
8.35 8.35 8.34 8.33 8.32 8.31 8.30
39.5 7 31.54 7 29.31 7 26.12 7 25.13 7 22.68 7 20.24 7 17.82
8.32 8.31 8.31 8.30 8.29 8.28 8.27
40.0 7 35.70 7 33.47 7 30.27 7 29.28 7 26.83 7 24.38 7 21.96
8.29 8.28 8.27 8.27 8.26 8.24 8.23

40.5 7 39.84 7 37.61 7 34.41 7 33.42 7 30.95 7 28.50 7 26.07

8.25 8.25 8.24 8.23 8.22 8.21 8.20
41.0 7 43.97 7 41.73 7 38.53 7 37.53 7 35.06 7 32.61 7 30.17
8.22 8.21 8.20 8.20 8.19 8.18 8.17
41.5 7 48.08 7 45.84 7 42.63 7 ' 1.63 7 39.16 7 36.70 7 34.26
8.18 8.18 8.17 8.17 8.15 8 . 14 8.13
42.0 7 52.17 7 49.93 7 46.72 7 45.72 7 43.24 7 40·77 7 38.32
8 . 15 8.15 8.14 8.13 8.12 8.11 8.10
42.5 7 56.25 7 54.00 7 50.78 7 49.78 7 47.30 7 44.82 7 42.37
8.12 8.11 8.10 8 . 10 8.09 8.07 8.06

43. 0 8 0.31 7 58.06 7 54.83 7 53.83 7 51.34 7 48.86 7 46.40

8.08 8.08 8.07 8.06 8.05 8.04 8.03
43.5 8 4.35 8 2.09 7 58.87 7 57.86 7 55.36 7 52.88 7 50.41
8.05 8.04 8.03 8.03 8.02 8.00 7.99
44.0 8 8.37 8 6.12 8 2.88 8 1.88 7 59.37 7 56.88 7 54.41
8.01 8.00 8.00 7.99 7.98 7.97 7.96
44.5 8 12.37 8 10.12 8 6.88 8 5.87 8 3.36 8 0·87 7 58.39
7.97 7.97 7.96 7.96 7.94 7.93 7.92
45.0 8 16.36 8 14.10 8 10.86 8 9.85 8 7.33 8 4.83 8 2.35
7.94 7.93 7.92 7.92 7.91 7.90 7.88

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

Travel times for h = 0 s

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

250 40.2 39.8 38.7 1 09.8 1 10.3 1 07.5 29.6 1 22.8

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

SEISMOGRAPHS AND SEISMOLOGICAL OBSERVATORIES

6.1 GENERAL CONSIDERATIONS

Seismographs play a similar role for seismology to that which X-ray

machines play for medicine or telescopes play for astronomy. They make the
otherwise inaccessible parts of the Earth "visible" and available for detailed
investigations. In spite of the fact that earthquakes had already been treated
as natural phenomena by ancient Greek philosophers, e.g. by Aristode
(384-322 B.C.), seismology became an established branch of natural sciences
only after the first reliable seismographs were developed and deployed at the
end of the nineteenth century.

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.

Broadly speaking, a seismograph is an instrument which monitors and in a

certain form records ground vibrations (as a continuous function of time)
caused by arriving seismic waves together with a very precise timing
information. The record of the motion is called the seismogram. The core of
modern seismographs is the seismometer which transforms the energy of
arriving seismic waves into electrical voltage. It works as a seismic-to-electric
transducer (sensor, detector) and senses the displacement, velocity or
acceleration of the ground motion. Seismometers are usually oriented in such
a way that they measure one of the three components (e.g. east-west,
north-south or vertical) of the ground motion. Thus, to be able to restore the
complete picture of the translational motion due to seismic waves emerging
at the Earth's surface, many observatories operate combinations of three
seismometers (sometimes called the complete station) oriented in the three
directions.

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

or Grenet-Coulomb, the mass is supported against gravity by various helical

springs, while the recent Wielandt-Streckeisen seismometer, developed in the
mid-1970's (Wielandt, 1983), makes use of a leaf-spring suspension. Seismic
waves that reach the recording site cause the frame to move together with
the bedrock, whereas the mass, due to its inertia and loose coupling to the
frame, tends to remain stationary. The relative motion of the mass with
respect to the frame is magnified, by mechanical, mechanical-optical or
electromagnetic means or by electronic amplifiers, and recorded. Nowadays,
magnifications of the order of 10 and greater can easily be achieved by
6

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.

One of the significant instruments employed by seismologists, which is not

based on the pendulum principle, is the strain seismometer or strainmeter,
invented by H. Benioff in 1935. This instrument essentially measures a
relative displacement between two piers on the Earth's surface (or at shallow
depth) some 20 m apart. It consists of a horizontal steel or quartz tube
anchored to one pier and extending sufficiently close to the other pier. A
proper transducer transfers the varying gap between the free end of the tube
and the second pier into a measurable electric signal. In contrast to pendulum
seismometers, which monitor the ground displacement, velocity or
 57

acceleration, the strain seismometers measure a component of ground strain

or deformation and are sometimes (Lee, 1987) called deformation
seismometers.

Yet another type of non-pendulum sensor is the piezoelectric seismometer

also called pressure detector. This type of instrument is usually deployed
during special studies carried out in submerged areas. With respect to the
rather specific field of application, strainmeters and piezoelectric
seismometers will not be discussed further.

The final, but not less important, component of a seismograph, in addition

to the sensor and the recording unit, is an accurate clock. To determine
arrival times of various recorded phases, a precise time indication on the
seismogram is required. Many modern seismographic stations keep their clock
accuracy at 1-10 ms by making daily comparison with the radio time signals
transmitted by standard world time services such as e.g. WWV in the United
States or DCF77 in FRG. An exhaustive list of radio transmitted time signals
may be found in Willmore (1979). To provide for a common time basis,
seismologists all over the world use Universal Co-ordinanted Time (UTC)
which is the scientific equivalent of Greenwich Mean Time, used for civil
purposes in Great Britain.

6.2 RECORDING SYSTEMS

Recording of seismic waves can be accomplished in many different ways.

Older seismographs, e.g. Wiechert or Mainka, employ a purely mechanical
direct recording by which a trace or seismogram is produced on paper by a
delicate ink-pen or on a smoked paper by a solid stylus attached to the
pendulum. Mechanical direct-recording instruments provide low magnification.
Some other older instruments, such as e.g. Milne-Shaw or Wood-Anderson,
make use of mechanical-optical recordings. A mirror which is fastened to the
pendulum or other moving part produce the trace by reflecting a light beam
on photographic paper.

More modern instruments use electromagnetic or, to some lesser extent,

electrostatic recording. In the former a voltage is generated either by the
displacement of a coil in a constant magnetic field (moving-coil
seismometers like e.g. Galitzin or Grenet-Coulomb) or by variations of the
magnetic field enclosed by a coil (variable-reluctance seismometers such as
e.g. Benioff). In both cases, the induced electromagnetic force is proportional
to the time derivative of ground displacement. This means that within a
certain period range, the recorded trace is proportional to the ground velocity.
In 1906, B.B. Galitzin coupled the seismometer generated electrical voltage to
a galvanometer and recorded its deflection optically on photographic paper by
58

means of galvanometer mirrors. Galvanometric recording is still in use at

most of the world's observatories. The paper is attached to a slowly rotating
drum which at the same time advances along its rotational axis (helical
motion). This arrangement allows for 24-hour, or even longer, uninterrupted
recording intervals between successive paper changes. The disadvantage of
recording on photographic paper is that the recording part of the system must
be located in a dark room which means that the recorded trace cannot be
examined immediately. One usually records one day of data on one sheet of
paper of the size approximately 30 χ 90 cm. The resolution often is 1 hr/line
and 24 lines/sheet or 15 min/line and 96 lines/sheet for long- and
short-period recordings, respectively. Hence, once the paper is developed and
fixed it can be quickly scanned for evidence of seismic activity during an
interval of 24 hours. A secondary advantage is the alignment of time marks
across the record which enables easy and rapid time estimates (hours and
minutes).

The electrostatic seismometers or variable-capacitance seismometers make

use of two systems of capacitor plates. One system is connected to the
pendulum arm while the other system is attached to the frame. Relative
motion of the capacitor plates, which are part of an electronic circuit,
modulates an electrical signal which is then applied to a sensitive
galvanometer for recording as in the preceding arrangement. When compared
with the electromagnetic recording, the variable-capacitance instruments are
nowadays rather scarce. However, it is worth mentioning that electrostatic
seismometers were the instruments which were installed and operated on the
Moon between 1969 and 1977 (five Apollo missions). Also some of the
operating strain seismometers employ this type or recording.

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

6.3 INSTRUMENTAL FREQUENCY CHARACTERISTICS

Pendulums have a special property, namely that after being excited by an

arbitrary motion of the frame and subsequently left alone, they swing at a
characteristic frequency, /„ (or period 7 J , called the natural frequency,
eigenfrequency (from German eigen meaning own) or free frequency.
Within certain limits, /„ is amplitude invariable. It is obvious that this
property has a direct consequence for the behaviour of pendulum
seismometers. When the ground moves forth and back very slowly at a
frequency much smaller than f , the pendulum's mass follows exactly the
H

movement of the ground. There is no relative movement between the mass

and the frame. In other words, when the frequency, / , of the Earth vibrations
decreases towards very small values, the amplitude, F , or the relative
movement (pendulum vs frame) tends towards zero. If the frequency /
increases, the amplitude F also increases and at the resonant frequency, /„,
the amplitude F reaches its maximum value. If we further increase the
frequency / above f the pendulum's mass lags behind the ground vibrations
H

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

Fig. 28. Amplitude, F, of a freely

coupled pendulum as a function of
ground-motion frequency, / . The
pendulum has an eigenfrequency /„.
60

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.

For seismometer-galvanometer systems, the relative movement between the

mass and the ground becomes more complicated because galvanometers have
oscillatory characteristics of their own. A seismometer-galvanometer circuit
operates as a feedback system so that not only the seismometer but also the
galvanometer influences the entire appearance of the seismogram. The
frequency dependent sensitivity of a seismometer-galvanometer system is
called the magnification characteristic, frequency characteristic or
response characteristic of the system. By a proper choice of free periods
and other instrumental constants (damping, coupling) for the seismometer and
galvanometer, nearly optimum shapes of magnification curves for many
particular purposes may be achieved. For example, systems may be designed
with practically constant characteristics in a wide frequency range or very
short and very long periods may be suppressed and specific portions of the
spectrum may be enhanced on seismograms. Several examples of different
frequency characteristics are shown in Fig. 29. Seismometer and galvanometer
free periods as well as maximum magnifications for several of the
instruments which remain in fairly common use are summarized in Table 2.

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

teleseismic events cover approximately the same amplitude range.

It is quite clear that a single instrument of the type already discussed

cannot cover the whole spectrum of frequencies and the wide dynamic range
generated by various seismic sources. In an attempt to overcome this
drawback, two or more sets of instruments with different frequency
characteristics and magnification levels are often installed side-by-side at
multi-purpose seismological observatories. For example, as follows from Fig.
29, by combining the SP and LP WWSSN systems, we can enhance seismic
signals with periods around 1 and 20 s and at the same time we can filter
out the undesired microseisms with dominant periods around 6 s. Hence,
operating different systems in parallel is clearly superior, although more
expensive, to one-system recording.

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

Instrument type Free period Magnification

Seismometer Galvanometer
T (s)
t T (s)
g

Benioff, SP (WWSSN) 1
1.0 0.7
Grenet-Coulomb 1.4 0.7 -10 4

SKM-3 1.6 0.4 2 χ 10 4

flat for Τ < 0.8 s

Willmore 1.0 0.3
S-13 0.75-1.1
Wiechert, 1000 kg 10 - 2 0 0 at 10 s
Milne-Shaw 10 - 3 0 0 at 10 s
SK 10-25 -1.0 constant 1-10 s
Benioff, LP 1.0 90
Press-Ewing (WWSSN) 1
15 100 -10 3

Galitzin 10-15 10-15 -10 3

Kirnos 22 80 -10 3

1
WWSSN stands for World Wide Standardized Seismograph Network

Instruments currently in use usually belong to one of the three following

categories (see also Fig. 29):
- Short-period (SP) instruments having the peak magnification in the period
range 0.1 - 1 s for detection of all seismic waves from local and regional
events and teleseismic body waves. The maximum magnification, which
obviously is site dependent, is usually of the order of 10 or 10 . 4 5

- Long-period (LP) instruments with peak magnification at periods of 10 s

or larger for detection of mainly surface waves. The maximum magnification
is usually around 5 χ 10 . 3

- Broad-band (BB) instruments, which record, with nearly constant magnifica-

tion, a broad band of periods from a fraction of a second to several
hundreds of seconds.

It is worth emphasizing that the appearance of an earthquake on the

seismogram is much influenced by the shape of the magnification
characteristics of the seismograph. Two or more systems with different
response curves, recording one and the same earthquake at a given
seismographic station, will produce seismograms which may be as different
as night and day. For this reason, the analyst should not be surprised when
one of the systems clearly recorded a sequence of seismic waves whereas the
other system sometimes may not show a sign of arriving seismic energy. To
illustrate the instrument effect, in Fig. 30 there are displayed
vertical-component seismograms, of an earthquake in the Sea of Japan
(m=5.9, h=51l km), recorded at an epicentral distance of 75° by four
different seismographs (computer simulations). The striking differences among
Fig. 30. Computer simulations of seismograms from four different seismograph systems. The deep-focus earthquake (h = 570 km) in Sea of Japan on
March 9, 1977 (magnitude m = 5.9) is here used to demonstrate the influence of the pass-band width of a recording system upon the complexity of

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.

Another example, which makes use of the Romania earthquake of March 4,

1977, is presented in Fig. 31. The bottom trace is an analog display of a
digital vertical-component, broad-band, velocity record made at Grafenberg,

Fig. 31. The Romania intermediate-

focus (h = 112 km) earthquake of
March 4, 1977 (magnitude m = 6.1)
recorded at Grafenberg, Bayern, FRG,
at an epicentral distance of 11°. The
exhibited traces are (from top to
bottom): 1) computer simulation of a
short-period WWSSN record; 2)
computer simulation of a long-period
WWSSN record; 3) analog display of
an actual broad-band digital recording.
All traces show the vertical-
component of the ground motion
velocity. For more details see the text.
After D. Seidl (personal
communication, 1988).
 65

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.

6.4 ANALOG AND DIGITAL RECORDS

Solutions to many seismological problems demand seismic records from a

number of stations. Therefore, further progress in earthquake seismology
relies upon continuing international cooperation including first of all a rapid
exchange of primary observational data (seismograms) among the existing
stations/observatories. During the last one hundred years or so, many different
types of seismographs have been successively installed at various sites around
the world. Some of them have been in operation since the turn of the
century, but many have been deployed after 1950, e.g. the approximately 120
seismographic stations distributed in 60 countries and equipped with identical
instrumentation constituting the World Wide Standardized Seismograph
Network (WWSSN). Already around 1920, the world-wide system of
earthquake observatories comprised more than one hundred permanent
stations. Today, there are approximately 1500 seismographic stations
operating around the world and involved in recording, analysis and
international data exchange. The majority of these stations still produce
seismograms in a rather traditional analog form often on sheets of
photographic paper. The available collection of millions of photographic
records stored at many observatory archives as well as at several international
data centers and the continuing flow of new seismograms produced every
day, obviously represent an invaluable source of information frequently
employed in routine seismogram analysis and in research.

The advent of fast digital computers and their massive application in

solving a broad spectrum of geophysical problems accelerated the
development of digital recording on magnetic tapes, floppy discs, etc. By
digital recording we here mean a system essentially consisting of a
seismometer generating an analog electric signal and an analog-to-digital
66

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.

A direct access to digital recordings facilitates seismogram analysis

enormously. The possibility of expanding the interesting portions of the
record, of filtering out the undesired frequency components, of displaying the
wave particle motion, of rotating the seismograph axis to any orientation, of
making changes in the response characteristics, etc, represent invaluable tools
to reveal trace details and information otherwise undiscernible on the record.
To illustrate the advantages in more detail, let us have a look at analog
displays of digital records presented in Fig. 32. Records of two large
earthquakes in Michoacan, Mexico, made at Matsushiro, Japan, at an
epicentral distance of 101° are shown. The instrument used is a long-period
vertical-component seismograph. The upper trace exhibits a record of the
main shock which occurred on September 19, 1985 at a depth of 28 km
(M=8.1). The lower trace shows the largest aftershock (M=7.5) which took
place on September 21, 1985 at a depth of 42 km. Both traces are
accompanied by enlarged initial portions, 6 min or so, of the original record.
Note that the first two cycles of Ρ from the two shocks have a similar shape
except for a new onset in the record of the main shock, following about 30 s
after P, which in its turn increases the width of the second cycle. The second
onset which is attributed to the complexity of the source, would certainly
pass unnoticed if the interpretation had been done on the original (i.e. not
enlarged) record alone.

Traces (computer plots) displayed in Fig. 33 serve to illustrate another

advantage of digital records which make it possible to enhance certain
frequencies during the interpretation while other spectral bands are efficiently
suppressed. The top trace exhibits an analog display of the original digital
record of a microearthquake recorded at a distance of about 80 km. There is
a distinct Pg onset, however it is difficult to discern the Sg arrival. The
identification of Sg has been done in two steps. First, all frequencies above
12 Hz were filtered out (middle trace). The filtering makes the Sg group
much more distinct when compared with the original record (upper trace).
Second, within the time window covering the Sg wavetrain the time scale
was extended by a factor of about 4 (bottom trace). The Sg arrival becomes
very distinct and can be measured with a desired accuracy.

Still another advantage of digital recordings on magnetic media is the

possibility of efficient data decimation. Recorded seismic events take on
average only a small fraction, 5% or so, of the total recording time. Thus,
 67

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

This chapter comprises a collection of 55 plates with actual seismograms.

Included are also phase interpretation and commentaries of specific features
reflected in the displayed records. Whenever possible, three-component
recordings are exhibited. The channel orientation, Ε for east-west, Ν for
north-south and Ζ for vertical, is given in the upper right-hand corner of the
shown records. If only one component is displayed, then no specification is
made in the seismogram, but the component is noted in the corresponding
plate caption. Collected seismograms were made at a number of
seismographic stations by a variety of instruments covering different
frequency bands. Categories of used instruments (short-, medium-, long-
period, broad-band) again are given in respective plate captions. In all plates,
time advances from left to right and, if not mentioned otherwise, there is one
minute between successive time marks (small upward displacements).

Exhibited seismograms cover the whole range of epicentral distances, a

large range of focal depths and various types of propagation paths. Different
seismic sources like tectonic and volcanic earthquakes, underground nuclear
explosions and several more exotic sources (sonic booms, collapses after
underground nuclear explosions) are presented. More classical analog records
as well as analog displays of modern digital records are displayed. The
arrangement of plates is according to epicentral distance except for records
from volcanic earthquakes and exotic sources which follow the suite of local
and regional events. The phase nomenclature used in the plates is naturally
the same as that used in Chapters 4-6. For the ease of the reader, the
notations of seismic waves used in the following plates are listed in Table 3
in the order as they appear in the text.

Source parameters like date, origin time, geographical region, epicentral

coordinates, focal depth and magnitude of events used are listed in Table 4.
If not mentioned otherwise the information is taken from ISC bulletins. For a
number of events recorded at local or regional distances, Table 4 was
completed with data provided by workers who interpreted the seismograms.
Observe that in some cases there is a certain disagreement between source
parameters (source region determinations, focal depth estimates) given by ISC
(see Table 4) and by seismogram interpreters who contributed to this book
(see plate captions). Permanent seismographic stations employed are listed in
Table 5. The table lists station codes and names, geographic regions and
coordinates. The stations are listed alphabetically by station code.
70

TABLE 3
Nomenclature of seismic phases

Symbol Meaning

Local and regional events

P, S Direct compressional or shear wave traveling through the upper crust

(observed only at very short epicentral distances).
Pg.Sg Compressional or shear wave in the granitic layer of the crust.
PmP, SmS Compressional or shear wave reflected at Moho.
Pn, Sn Compressional or shear wave traveling along (just beneath) the Moho
discontinuity, so called head wave.
P*, 5* (or Pb, Sb) Compressional or shear wave traveling along (just beneath) the Conrad
discontinuity.
pPn Depth phase that leaves the focus upward as P, is reflected as Ρ at the free
surface and continues further as Pn.
sPn Depth phase that leaves the focus upward as 5, is reflected and converted
into Ρ at the free surface and continues further as Pn.

Rg Short-period crustal surface wave of Rayleigh type.

Lg Guided crustal wave traversing large distances along continental paths.
Τ Compressional wave propagating through the ocean (tertiary wave).
Τ phases are occasionally observed even at large teleseismic distances.
TPg (TSg, TRg) Wave that travels the ocean and land portion of the total transmission path
as Τ and Pg (Sg, Rg), respectively.

Teleseismic events

P, S Direct compressional or shear wave, so called elementary or main wave.

PP, PPP, SS, SSS Ρ or S wave reflected once or twice at the Earth's surface.
SP S wave converted into Ρ upon reflection at the Earth's surface.
PPS, PSP, PSS Ρ wave twice reflected/converted at the Earth's surface.
PcP, ScS Ρ or S wave reflected at the core-mantle boundary.
PcS, ScP Ρ or S wave converted respectively into S or Ρ upon reflection at the core
mantle boundary.
Depth phase that leaves the focus upward as Ρ (ρ leg), is
pP, pS, pPP, pPS etc reflected/converted at the free surface and continues further as P, S, PP, PS
etc.
Depth phase that leaves the focus upward as S (s leg), is
sP, sS, sPP, sPS etc reflected/converted at the free surface and continues further as P, S, PP, PS
etc.
Ρ wave reflected at the underside of a discontinuity at depth d in the upper
PdP
part of the Earth, d is given in kilometers, e.g P400P.
Pc, Sc (or Pdif, Sdif) Ρ or S wave that is diffracted around the core-mantle boundary.
PKP (or P') Ρ wave traversing the outer core.
PKS Ρ wave converted into S on refraction when leaving the core.
SKS S wave traversing the outer core as Ρ and converted back into S when
again entering the mantle.
 71

TABLE 3 (continued)

Symbol Meaning

SKP S wave converted into Ρ on refraction into the outer core.

PKP1, PKP2 (or PKP , BC
Different branches of PKP.
ΡΚΡΛΗ)

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.

LQ Surface wave of Love type.

G Mantle wave of Love type.
R Mantle wave of Rayleigh type.
LQ-type mantle wave that travels the direct and anticenter routes. Waves
which have, in addition, traveled once or several times around the Earth
Gl, G2
are denoted G3, G4, G5, G6 etc.
L/?-type mantle wave that travels the direct and anticenter routes. Waves
that have, in addition, traveled once or several times around the Earth are
Rl, R2
denoted R3, R4, R5, R6 etc.
 72
TABLE 4
Source parameters of events used

Date Origin Geographical Epicentral Focal Magnitude 1

Comments 2

time region coordinates depth

y m d h m s km

52 11 04 16 58 22 Kamchatka 52.6 °N 160.3 Έ Μ = 8.25

59 06 27 19 11 29.3 China-USSR 41.87°N 80.01Έ 27
border region
62 03 18 15 30 30 Albania 40.70°N 19.65Έ 13
63 01 19 19 54 08 South-central 59.8 °N 15.1 Έ 1-2 Rockburst. Uppsala solution.
Sweden
64 10 21 23 09 19.0 Indian-China 24.04°N 93.75Έ 37 m = 5.9
border region
65 11 13 04 33 50.6 Northern Sinkiang 43.87°N 87.74Έ 29 m = 6.2
Province, China
67 11 21 17 02 25.8 Norwegian Sea 72.66°N 8.14Έ 33 Μ = 5.4
68 08 10 02 07 00 Moluca Passage 1.38°N 126.24Έ 1 m = 6.3
68 08 14 22 14 20.1 Northern Celebes 0.06°N 119.73Έ 22 m = 6.1
69 03 31 19 25 27.0 Sea of Japan 38.49°N 134.52Έ 397 m = 5.7
69 05 05 05 34 24.4 North Atlantic 35.99°N 10.34°W 37 m = 5.5
Ocean
69 06 12 15 13 30.9 Crete 34.43°N 25.04Έ 22 m = 5.8
70 10 14 05 59 57.3 Novaya Zemlya 73.3ΓΝ 54.89Έ 0 m = 6.6 Underground explosion.
71 07 26 01 23 21.2 New Ireland 4.93°S 153.18Έ 43 m = 6.6
region
72 06 11 16 41 02.7 Celebes Sea 3.86°N 124.26Έ 336 m = 6.2
73 09 15 01 45 57.9 Iceland region 63.86°N 22.23°W 1 m = 5.3
74 08 08 01 25 13.9 Greenland Sea 73.19°N 6.3 Έ 20 m = 5.0
74 11 02 05 50 30.2 Novaya Zemlya 70.6 °N 54.1 Έ m = 7.0 Underground explosion.
Μ = 6.2 Uppsala magnitudes.
75 05 16 03 01 01 Sweden 60.1 °N 15.0 Έ 1-2 M = 2.6
L Rockburst. Uppsala magnitude.
75 08 13 16 29 19 South-central 59.8 °N 15.1 Έ 1-2 M = 2.8
L Rockburst. Uppsala solution.
Sweden
76 07 27 00 56 43.4 Norwegian Sea 72.13°N 0.80Έ 9 m = 5.1
76 09 15 09 21 18.6 Austria 46.34°N 13.12Έ 12 m = 5.4
77 03 04 19 21 54.1 Romania 45.83°N 26.72Έ 86 m = 6.1
77 03 09 14 27 56.2 Sea of Japan 41.66°N 131.05Έ 556 m = 5.9
78 05 23 23 34 11.4 Greece 40.73°N 23.25Έ 9 m = 5.6
Μ = 5.8
78 09 03 05 08 30.3 Germany 48.30°N 8.93Έ 10 m = 4.9
Μ = 5.3
79 09 19 21 35 36.8 Central Italy 42.80°N 13.04Έ 6 m = 5.8
Μ = 5.9
80 05 18 15 32 14.3 Washington State, 46.38°N 122.02°W 4 m = 4.6 Mount St. Helens eruption
USA Μ = 5.2
80 07 16 19 56 47.1 New Guinea 4.45°S 143.53Έ 87 m = 6.3
Μ = 6.9
81 01 02 07 37 00.5 South of Panama 2.13°N 79.16°W 27 m = 5.7
Μ = 5.8
81 01 10 09 47 24 South of Honshu 33.7 °N 139 Έ 40 M = 3.3
L Magnitude according to
M. Yamamoto (p.c.)
81 05 02 16 04 54.6 Afghanistan-USSR 36.40°N 71.15Έ 217 m = 5.9
border region
82 04 26 Asama volcano, 36.4 °N 138.6 Έ Volcanic earthquake. After
Japan M.Yamamoto (p.c.)
82 07 01 07 41 53.7 Andreanof Islands 51.39°N 179.94°W 51 m = 6.3
Μ = 5.5
83 02 05 Asama volcano, 36.4 °N 138.6 Έ Volcanic earthquake. After
Japan M. Yamamoto (p.c)
83 03 21 Asama volcano, 36.4 °N 138.6 Έ Volcanic earthquake. After
Japan M. Yamamoto (p.c.)
83 04 03 02 50 02.8 Costa Rica 8.80°N 83.11°W 44 m = 6.3

73
Μ = 7.2
TABLE 4 (continued)

74
Date Origin Geographical Epicentral Focal Magnitude 1
Comments 2

time region coordinates depth

y m d h m s km

83 09 12 15 42 08.3 Afghanistan-USSR 36.52°N 71.10Έ 205 m = 5.9

border region
84 02 01 07 28 28.4 Sea of Okhotsk 49.05°N 146.63Έ 568 m = 5.8
84 03 06 02 17 21.0 South of Honshu 29.35°N 138.92Έ 454 m = 6.1
84 04 20 06 31 10.3 Sea of Okhotsk 50.06°N 148.77Έ 578 m = 5.9
84 07 14 01 09 10.5 Eastern Kazakh, 49.85°N 78.92Έ m = 6.2 Underground explosion.
USSR Μ = 4.8
84 09 13 23 48 51.2 Honshu 35.78°N 137.49Έ 18 m = 6.0
Μ = 6.4
84 10 14-15 Near Izu Peninsula, shallow Earthquake swarm. After
Japan M. Yamamoto (p.c).
84 11 17 06 49 31.0 Northern Sumatera 0.22°N 98.05Έ 42 m = 6.2
Μ = 7.3
84 11 20 08 15 17.8 Mindanao 5.13°N 125.17Έ 218 m = 6.4
85 03 19 11 24 11.5 South Island, 44.99°S 167.31Έ 133 M = 4.0
L Magnitude according to
New Zealand M.E. Reyners (p.c).
85 03 23 12 00 26 Off west coast of 45.07°S 166.5Έ 4 M = 3.7
L Magnitude according to
South Island, M.E. Reyners (p.c).
New Zealand
85 04 20 19 35 34 Near south coast of 34.94°N 133.46Έ 8 M = 4.9
L Matsushiro magnitude. Poorly
Southern Honshu determined source parameters.
85 05 16 14 20 28 Mid-Indian Rise 29.04°S 77.78Έ 30 m = 5.9
Μ = 6.0
85 06 06 Asama volcano, 36.4 °N 138.6 Έ Vocanic earthquake. After
Japan M. Yamamoto (p.c).
85 06 15 00 40 19.3 Sweden 56.51°N 12.20Έ 33 M -- = 4.6
L Uppsala magnitude.
85 06 30 Asama volcano 36.4 °N 138.6 Έ Volcanic earthquake. After
Japan M. Yamamoato (p.c).
85 07 07 00 08 56.4 Switzerland 47.00°N 7.69Έ 10 M -- = 2.7
L Magnitude according to
D. Mayer-Rosa (p.c).
85 08 15 04 28 47.4 Northwestern 47.06°N 18.01Έ 10 m - 4.7
Hungary Μ = 5.0
85 09 19 13 17 50.1 Michoacan, Mexico 18.54°N 102.32°W 29 m = 6.4
Μ = 7.9
85 09 21 01 37 15.1 Near coast of 17.81°N101.69°W 42 m = 6.2
Guerrero, Mexico Μ = 7.5
85 09 25 07 43 57.3 Michoacan, Mexico 18.19°N102.81°W 30 m - 5.3
Μ = 5.2
85 10 09 09 33 32.6 South of Alaska 54.73°N159.65°W 30 m = 6.1
Μ = 6.5
85 10 26 16 34 58.9 Tuamotu Archipelago 21.81°S 139.0 °W m = 5.3 Underground explosion.
region
85 11 16 11 30 45.0 Near west coast 37.08°N 136.51Έ 313 m = 4.8
of Honshu
86 01 21 19 04 05.3 Germany 50.22°N 12.48Έ 10 M -- = 1.4
L Magnitude according to
A. PleSinger (p.c).
86 04 30 07 07 19.6 Near coast of 18.37°N130.01°W 38 m - 6.1
Michoacan, Mexico Μ = 6.9
86 05 01 13 27 56.1 Peru-Brazil 9.20°S 71.23°W 600 m = 6.0 After G.L. Choy (p.c).
border region
86 05 15 Volcano Rincon 10.8 °N 85.4 °W Volcanic tremors. After
de la Vieja, J.M. Protti (p.c).
Costa Rica
86 06 08 04 25 06.8 Off east coast 36.56°N 142.61Έ 35 m = 4.9
of Honshu Μ = 5.5
86 07 11 12 05 09 Near Matsushiro, 36.52°N 138.15Έ 6.9 M =: 1.8
L After M. Yamamoto (p.c).
Japan
86 08 20 10 09 32 Tonga 16.94°S 173.49°W 26 m = 5.5

75
 76
TABLE 4 (continued)

Date Origin Geographical Epicentral Focal Magnitude 1

Comments 2

time region coordinates depth

y m d h m s km

86 10 09 10 08 53.2 Switzerland 46.38°N 7.45Έ 6 M = 3.4

L Magnitude according to
D. Mayer-Rosa (p.c).
86 10 20 06 46 09.9 Kermadec Islands 28.13°S 176.40°W 29 m = 6.7
region Μ = 7.9
86 11 14 16 00 00.0 Southern Nevada 37.1 °N 116.0 °W m = 5.8 Underground explosion. After
Μ = 4.5 NEIC.
87 01 26 Pacaya volcano, 14.37°N 90.60°W Volcanic earthquakes. After
Guatemala E. Molina (p.c).
87 03 27 Arenal volcano, 10.47°N 84.68°W Volcanic tremors. After
Costa Rica J.M. Protti (p.c).
87 04 29 14 27 36.6 Fiji Islands region 18.93°S 177.86°W 392 m = 5.9 After G.L. Choy (p.c).
87 05 27 Pacaya volcano, 14.37°N 90.60°W Volcanic earthquakes. After
Guatemala E. Molina (p.c).
87 06 05 05 00 00 Southern Xinjiang m = 6.8 Probably underground explosion.
Province, China Uppsala solution.
87 07 27 09 00 49 Central Japan 33.02°N 140.17Έ 156 After M. Yamamoto (p.c).
88 04 27 Poas volcano, 10.18°N 84.25°W Volcanic earthquake. After
Costa Rica J.M. Protti (p.c).
88 05 30 Tacana volcano, 15.13°N 92.11°W Volcanic earthquake. After
Guatemala E. Molina (p.c).
89 04 20 Eastern Siberia m - 6.4 Uppsala preliminary solution.
Μ = 6.4

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

Code Station name Region Geographical

coordinates

ACR Cerro Adams Costa Rica 8.65°N 83.17°W

ANMO Albuquerque New Mexico, USA 34.95°N 106.46°W
AVOE Asama Volcanic Nagano, Honshu, 36.40°N 138.57Έ
Observatory Japan
BKS Berkeley California, USA 37.88°N 122.24°W
CHTO Chiang Mai Thailand 18.81°N 98.94Έ
COL College Outpost Central Alaska, USA 64.90°N 147.79°W
CYZ Clyde South Island, 45.15°S 169.33Έ
New Zealand
GOT Goteborg Sweden 57.70°N 11.98Έ
GRF Grafenberg array Bayern, FRG 49.69°N 11.22Έ
GRFO Grafenberg Bayern, FRG 49.69°N 11.22Έ
JAS Jamestown California, USA 37.95°N 120.44°W
KHC KaSperske Hory Czechoslovakia 49.13°N 13.58Έ
KIR Kiruna Sweden 67.84°N 20.42Έ
KTJ Kamata Shizuoka, Honshu, 34.94°N 139.09Έ
Japan
LLS Linthal-Limmern Switzerland 46.87°N 9.00Έ
LON Longmire Washington, USA 46.75°N 121.81°W
MAT Matsushiro Nagano, Honshu, 36.54°N 138.21Έ
Japan
MHC Mount Hamilton California, USA 37.34°N 121.64°W
MNA Mina Nevada, USA 38.43°N 118.16°W
MUO Muotatal Switzerland 46.98°N 8.67Έ
NUE Niue Cook Islands 19.08°S 169.93°W
POA Poas Costa Rica 10.15-N 84.22°W
PRI Priest California, USA 36.14°N 120.67°W
RAR Rarotonga Cook Islands 21.21°S 159.77°W
RIN Rincon de la Vieja Costa Rica 10.77°N 85.36°W
RSCP Cumberland Plateau Tennessee, USA 35.60°N 85.59°W
RSNT Yellowknife Northwest Territories, 62.48°N 114.59°W
Canada
RSNY Adirondack New York, USA 44.55°N 74.53°W
RSON Red Lake Ontario, Canada 50.86°N 93.70°W
RSSD Black Hills South Dakota, USA 44.12°N 104.04°W
SAO San Andreas California, USA 36.77°N 121.45°W
Geological Observatory
SBG Sibinal Guatemala 15.13°N 92.05°W

SCP State College Pennsylvania, USA 40.80°N 77.87°W

SLE Schleitheim Switzerland 47.77°N 8.49Έ

78

TABLE 5 (continued)

Code Station name Region Geographical

coordinates

TK02 Tokai O.B.S. Tokai, Honshu, Japan 33.95°N 137.76Έ

TMA Tamaro Switzerland 46.11°N 8.87Έ
TLO Toledo Spain 39.88°N 4.05°W
UDD Uddeholm Sweden 60.09°N 13.61Έ
UME Umea Sweden 63.83°N 20.24Έ
UPP Uppsala Sweden 59.86°N 17.63Έ
 79-80

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

6.00 7.00 8.00 9.00 10.Oü

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 5. Near-surface events (rockbursts) in south-central Sweden recorded by vertical seismographs at

epicentral distances between 76 and 300 km. From top to bottom: a) rockburst of May 16, 1975, A/L=2.6,
recorded at Uddeholm, Δ=76 km; b) rockburst of August 13, 1975, AfL=2.8, recorded at Uddeholm, Δ=76
km; c) rockburst of January 19, 1963, magnitude unknown, recorded at Uppsala, Δ=148 km; d) rockburst
of January 19, 1963, recorded at Göteborg, Δ=300 km. All four records exhibit remarkably well developed
short-period Rg waves with normal dispersion within a period range from 1 to 0.5 s (top two traces) and
from 3 to 1 s (bottom two traces). Rg dominates the records emphasizing the shallow focus depth, of
probably not more than 1-2 km. Traces a) and c) display also clear Pg and Sg.
Plate 6. Short-period three-component records from a nearby large shallow earthquake. The recording site is
Matsushiro, Japan. This earthquake (Λί=6.4) occured in central Japan on September 13, 1984 at a depth of
14 km and at a distance of 99 km from Matsushiro. The recorded earthquake is interpreted as a double
event, small and large, about 2.4 s apart. The multiple Pg phases, denoted Pgl and Pg2, are clearly visible
in all three channels. The corresponding multiplicity in Sg can be seen, though somewhat less distinctly, in
the horizontal channels. There are 20 s between successive time marks at the bottom of the plate.
 DEPTH IN KM
I I I I I I REDUCED TIME IN SEC l-D/B.O (s)
(J) Ü1 ►C O) (\) - i »— ·— ru no
CD CD CD CD CD CD CD en o en o en o en
 87

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

Plate 8. Short-period vertical-component seismograms of a shallow (A) and intermediate-focus (B)

earthquake in the Fiordland region of New Zealand made at Clyde, New Zealand. Source parameters of the
two events are the following:

Event Date Ml Epicentral coordinates Focal Epicentral

depth distance

A Mar 23, 1985 3.7 45.16°S 166.46Έ 12 km 225 km

Β Mar 19, 1985 4.0 45.09°S 167.50Έ 106 km 144 km

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:

Event Date ML Epicentral coordinates Focal Epicentral

depth distance

A June 8, 1986 5.4 36.53°N 142.96Έ 35 km 426 km

Β April 20, 1985 4.9 37.41°N 142.77Έ 35 km 417 km

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 17. Examples of short-period vertical-component seismograms from explosion-type volcanic

earthquakes. Top and middle traces show two shocks from the Arenal volcano, Costa Rica, on March 27
and 28, 1987, respectively. Both events were recorded at station ACR located 3 km from the crater. The
bottom trace is an explosion-type earthquake from the Asama volcano, Japan, recorded at the Asama
Volcanic Observatory on April 26, 1982 at a distance of 3.2 km from the crater. As follows from the
displayed traces, seismic signals generated by explosive eruptions have longer periods and longer duration
than signals from Α-type events.
98

Plate 18. Examples of short-period vertical-component seismograms from explosion-type volcanic

earthquakes generated by the Pacaya volcano, Guatemala. Displayed records were made on May 27, 1987
and on January 26, 1987 by portable seismograph stations deployed 1 km (top traces) and 4 km (bottom
traces) from the crater, respectively. Note that the recording speed for the bottom traces was twice as large
as for the top traces.
 99

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

**\Arv WÂ W/Y^^^A^A/^^ ,«Λ^^A^^A^/>«^J^^^r^^^

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

-=zP'. :PP: üPcP=r

iL — a
ν'^^Ô^ÉÀÃ?!.^?^!-.,^-!.^.,..^'..^

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

Plate 26. Seismograms made at Uppsala,

Sweden, on short-period vertical-component
(top trace), medium-period three-component
(three middle traces) and long-period vertical-
component (bottom trace) seismographs. This
earthquake occurred on Crete, Greece, on
June 12, 1969 (M=6.2, A=25 km) at a
distance of 26° from Uppsala. The onset of P
is distinct on all the records. The free-surface
reflections, PP, and the core-mantle reflection,
PcP, are visible on the short-period record.
The S wave arrives about four and a half
minutes after P and is easily recognizable,
especially on the N-S component medium-
period and vertical long-period records.
Medium and long-period traces also show a
clear dispersive Rayleigh wave train, starting
2-3 minutes after S. The sudden increase of
Lfl-wave amplitudes at approximately 15:30
(medium and long-period records) is due to
the arriving Airy phase.
 111—112

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:

Î\JJ*I^^M ~U^nT> Ã ^ é é ^ ^ ^ ^ Ì ^ ^ I I > ^ ' ^^ww ^ ^ I W A ^ ^ ^ N J ^ ^ ' »

^"^*^^^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

Λ»**^*»**^*^^ Lûioceantc) w**w* LÛ(continental) *+«prtw*t^ ^Ï

*V**M*4I/W^^

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

—S- —G— -/ \.VÛ L·

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

Plate 32. Seismograms from the underground nuclear

•*«VVVVW^^M' test in Semipalatinsk area, eastern Kazakh, USSR
(m=7.0, Ë/=5.0), on July 14, 1984 made at Uppsala,
Sweden, at an epicentral distance of 35.5°. The upper
three traces show the short-period seismograms
whereas the lower three traces display long-period
records. The traces exhibit in a remarkable way a
NI series of features important for successful
discrimination of tectonic earthquakes against
éééËé m ii»,»»\^vs*v^*»vvVV^»*^^ underground nuclear tests: 1) A strong P wave is
seen in the short-period seismograms. The first onset
has always a compressional character, i.e. the first
swing on the vertical trace corresponds to the upward
ground motion. 2) Surface waves are much weaker
than one would expect from a shallow-focus event
LR" with a body-wave magnitude m=7.0. Note also the
large discrepancy between m and M. 3) S waves are
vA^l/vw^^ practically absent on the records. The event is located
to the east of Uppsala and hence P and LR waves
are best seen on the vertical and E-W components.
 123

><! !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—

l ^ ^ '« I »II -«•^V^\<»»*»^>1—»

« 1
• «^^V ^*^»*»

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**^**

:05··13=-: PkikP--- :Z(cont.)

TiyV^ W,, Λ , I
^^V^V•»V•S//V* »**^^^ ^*'-' ^'^ " ' "A*

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

Plate 38. Seismograms from the underground nuclear

explosion on June 5, 1987 in southern Xinjiang,
China (m=6.8), made at Uppsala, Sweden, at an
epicentral distance of 45.5°. The arrangement of
records is the same as that used in Plate 32. All
features typical for underground explosion records (cf
Plates 32 and 36) are well exhibited. Note especially
the absence of S waves on the records. The arrow on

=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

ï r ã*Λ" i-^^NL»w—»/»»^' —^'» ã» ^»v^/V" ' ô

_ _" w' v v^í^
'WW~-

^»Mwwwwyw^^^'^' 'Àéãé V WW m,m1ij »i »y^^A/V^/V^»^

*w y w W W ^ ^ w ^ » Vufi
m lATLrii-ij-i-i-uVLruVu
- ééÀ.
é 'jyijl·
' - ILI_IU
I .Àã\Λ/ííõ~õ
i
r
~ ''Wt'-~'-
' irii-uVn.
1 ij->nj-M-TJ^.i^i-»-u—i—
i <*|/u*-*M
* · , ' , · - - «- f V f l j'V' ««« <»»
. ^ »» . « . « / w»i / .» « A A A W « W A A A A W « < V I >V V > W W « i- y v V^tyV%*»s»^^^VA— . >r. w^«^
N
^^VV^*»-" V^l/l/VA/lA AA/l A/v^^

> l M , < VW1 •*AA/W«*< ^^^^.«>.-^y.—WVWv»—^-VN^^»»»

w v ^.%^<^-*^>^^\r' '' *«»V^VV WN/wy i ^ V ^ - · " «*» »V» W / V w ' ^ y V W ^ ' V W » ' | Wv^^WWWV^***' MA/WVI/ —WWv^'YN^A/W-V<— » ^ ^^«i-^^f -<^W/V^»
T
>Y/VWV»»>~««VV''¥*N^>^rVVA» *|fV*^' ^" ^'^

* > 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

Plate 39. Long-period records of an

earthquake in eastern Siberia (A/=6.4) on
April 20, 1989, made at Uppsala, Sweden.
These are typical seismograms from a
predominantly continental path. Direct and
surface-reflected body waves are clearly seen
on the records. The surface waves exhibit
pronounced continental-path dispersion
characteristics. The LQ wave train can be
observed in the N-S component (upper trace)
while LR are best seen in the vertical
component (lower trace). The vertical channel
shows both the fundamental and the higher-
mode Rayleigh wave groups. Unipolar pulses
seen at the bottom of both traces are
calibration pulses which provide quick checks
of the frequency characteristic of the entire
system. Calibration pulse is a response of the
system to a single electrical short impulse (or
step function) injected into the seismometer.
137—138

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

1.4E+03q 1.6E+04: SKS

.SNY RSNY i
>JSS 78.6 vVv*^|4^ NS 78.6 :
-1.4E+03 -1.6E + 04 :
2.5E+03 1.4E+04:
RSNY RSNY I
H *V*V*lfï|
l^^lfiN^^T^^ EW 78.6 :
EW 78.6 Plate 40. Seismograms form a deep-focus (A=588 km)
-2.5E+03 -1 1 1 1 1 1 1 1 1 1 r- -1.4E+04 : m 1 1 1 1 r earthquake in the Sea of Okhotsk on April 20, 1984 (m=5.9),
made at a suite of stations located at epicentral distances
between 50.1 and 82.2°. Short-period (left) and long-period
1.4E+04: SkS+S (right) horizontal-component record sections are exhibited.
7.0E+03q ScS [ScS Station codes, epicentral distances, component identifications
RSCP 1 sksë RSCP (NS or EW) and amplitude scalings are given to the left of
displayed traces. Broadly speaking, 5 waves are best
NS 82.2
*%· l^tf^Jt^^^i^w^rJpmiv'V^r*'^
NS 82.2 : recognized on long-period records. Note that in the long-
period records, the amplitude of various transverse phases are
-V.OE+OS11 -1.4E+04 = 2-10 times larger when compared with those on the short-
1.2E+04 3 2.5E+04- period records. Four distinct phases, namely 5, ScS, sS and
SS are identified in the long-period seismogram from RSNT.
RSCP , RSCP ! At distances larger than, say, 60° SKS will emerge. However,
EW 82.2 3 JtUf^gft^t/t^Jt.» #»*»» ^ i . t i ^ m » w > ^ M ^ >
EW 82.2 : in the present example, SKS observed at RSSD (Ä = 66.9°)
is mixed with ScS. S, ScS and SKS start to coalesce as
distance increases toward 82°. S, SKS and ScS are
-1.2E+04 1 1
1
r —1 1 1 1 1 r -2.5E+04 occasionally observed also on short-period recordings,
20 40 60 80 100 120 2 3 4 although their onset times are usually very emergent.
Time(s) Time (min)
 139—140

7.0E + 03: PcP P? 1.2E + 03 PcP pP PP sP

SCP I IJ
SCP
50.1
*öö#Ç^'^éΜÀ* *'"» *» '< 4< 50.1
-7.0E+03: -1.2E+03
4.0E + 04- PcP PP 1.4E + 04 PcP pP PP sP
RSNY : RSNY
53.6 ; ^ 4 ^ 4 » ^ I Nto »1» M* »4* mjm ««» » Λ*ì
53.6
-4.0E + 04: -1.4E+04
5.0E + 05- P PcP pP 1.0E+04
ANMO ! ANMO
55.1 : 55.1
-5.0E+051 -1.0E+04
3.0E+04- P
à PcP pP 1.2E+04
RSSD - RSSD
60.8 :
-3.0E+04J —L· 60.8
-1.2E+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—>

/v^Vty/^Vv^rv^YN^ Plate 43. Short-period (uppermost trace) and

long-period (lower three traces) seismograms
LIR«
15''51 from an earthquake in the Afghanistan-USSR
border region (m=5.9, A=209 km) on
September 12, 1983 made at Toledo, Spain,
at a distance of 57.5°. This is an interesting
example of a deep-focus, medium-size event.

^ 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

«^^ .m*»» V ^ V V -Airy phase

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

Plate 48. A deep-focus earthquake record

made at Jamestown, California, on a short-
period vertical-component seismograph. The
earthquake occurred on March 6, 1984
(m=6.1, A=454 km) south of Honshu, Japan,
at a distance of 78° from Jamestown. The
long separation of the pP and P phases
(101.5 s) and the absence of surface waves
indicate a deep focus. The wave train is

£
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

1SM Plate 54. Four earthquake records made at Toledo, Spain, on

a long-period vertical-component seismograph. The exhibited
records sample the epicentral distance from 115 to 140°. The
source parameters are the following:

Earthquake in Celebes on August 14, 1968 (Δ=115°, Λ/=6.0,

A=23 km).
Earthquake in Moluca on August 10, 1968 (Δ=119°, Λ/=6.3,
A=33 km).
Earthquake in Papua on July 16, 1980 (Δ=134°, Λί=6.5, Á=84
km).
Earthquake in New Ireland on July 26, 1971 (Δ=140°,
Λ/=6.3, Á=48 km).

Four examples of clear diffracted P (Pc) are shown. It

should be emphasized that these phases are not discernible
on corresponding short-period records. Among other features
it is seen that Pc exhibits longer periods when compared
with those of P'(PKIKP) and that the time interval between
P" and Pc clearly decreases with increasing distance. In
general, there is a large energy arriving between Pc and P'\
the physical cause of which is poorly understood. In some
cases, e.g. for the event on July 16, 1980, part of this energy
may be attributed to pPc. Pc and P" usually show the same
polarity, compression or dilatation. The most prominent
phase, in the whole P-wave group and the distance range
considered, is PP.
 165—168

5.0E + 02 PKlkP 4.0E + 02-

PKIKP
COL Ë*^ë' i^^M#Wff^W COL :
134.3 : 134.3 ·
-5.0Å + 02 : ! -4.0E + 0 2 J
8.0E + 0 3 3 P W » < P + ^ T > 1 + P i < P 2 2 . 5 E ^ 0 4 : PKIKP+PKP1 + PKP2
RSNT RSNT
145.6 ß^'^^éÇ^ 145.6
-8.¼Ε + 03 -2.5E + 04 «,,~Λ
4.0Ε + 03ç PKP1
PKP2 B.0E + 03q PKP1PKP2
RSNY A l
153.1 > ~ ^ M
RSNY
153.1
^ ^

-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.

lime es) Time (s)

 169

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

Numbers in boldface relate to Plate entries.

Aftershock, 5, 6, 38, 40, 66, 67 DCF77, 57

Airy phase, 43, 26, 44 Dehydration of minerals, 40
Anelastic attenuation, 16 Delta-shape envelope, 15
Angle of incidence, 2 1 , 24 Densification, 40
Angle of refraction, 21 Destructive plate margin, 3
Antipode, 34 Diffraction, 32
Appolo mission, 58 Dispersion, 15, 25, 32, 12, 23-26, 34, 36, 39
Asama Volcanic Observatory, 77, 15-17, 19 - see also Velocity dispersion
Asthenosphere, 3 - curve, 42, 43
-, inverse, 27, 43, 30, 50
b, 22 -, normal, 15, 16, 42, 5
B 34
y Dynamic range, 25, 61, 66
- see also Caustic point
Basaltic layer, 20 Earth contraction, 40
- rocks, 9 - cooling, 40
Body wave, 13, 38, 41, 47, 48, 50, 60 - model CAL6, 11
Butterworth low-pass filter, 68 Earthquake, 1
-, Α-type volcanic, 7, 15-17
c, 30, 42 -, Β-type volcanic, 7, 16
Calibration pulse, 39 -, collapse, 7, 8
Caustic point (Caustic), 33-37, 55 - - see also Implosion
Cavity collapse, 21 -, crustal, 15, 20, 25
Channel (guided) waves, 27 -, deep-focus, 6, 31, 38-40, 44, 64, 9, 11, 37,
- - see also Lg waves 40, 4 1 , 43, 45, 48, 52, 53
Clipped record, 25 -, explosion-type volcanic, 7, 17, 18
Coda, 13, 37, 38 -, fluid-injection induced, 8
- duration, 25 -, intermediate-focus, 6, 38, 40, 64, 8, 11
Complete station, 55 -, interplate, 5
Compressional (Longitudinal) wave, 13, 20 -, intraplate, 5
- - see also Ρ wave -, regional, 25, 60
Conrad discontinuity, 9, 22, 23, 54 -, reservoir induced, 8
Constructive plate margin, 3 -, shallow-focus, 6, 13, 16, 27, 38, 40, 44, 47,
Core, 9-11, 36 8, 25, 32, 42, 51
-, inner, 11, 30, 34-36 - source parameters, 1
-, liquid, 36, 37 - swarm, 6, 3
-, outer, 11, 30, 33, 34, 37 -, tectonic, 7, 8, 69, 32
- phase, 30, 33, 35 -, volcanic, 7, 8, 69, 73-76
Core-mantle boundary, 9-11, 30, 32, 33, 36 Edwards Air Force Base, 20
Coupling, 60 Energy release, 6
Critical angle, 21 - scatter, 19
- epicentral distance, 21 Epicenter, 4
Crust, 9, 10, 28 Event, 7, 8
continental, 10 -, controlled, 8
-, oceanic, 10, 23 -, double, 65, 2, 6
Cultural noise, 8, 17, 18 -, induced, 7
-, local, 20, 22, 23, 69, 70
d 31
y -, man-made, 8
Damping, 60 -, natural, 8
Data decimation, 66 -, near-surface, 24, 5, 2 1 , 24
172
-, regional, 20, 24, 69, 70 / , 35
- tape, 67 Japan Meteorological Agency (JMA), 38, 7
-, teleseismic (Teleseism), 20, 70 JB tables, 50-52, 54
-, triggered, 7
K, 33, 35-37
Fault plane, 4 Kernwelle, 33
- - see also Plane of weakness
Feedback, 60 Laboratory model measurements, 60
Floppy disc, 65 Leaf-spring suspension, 56
Focal depth, 38 Lg wave, 24, 25, 70, 27, 33
Focus, 4, 20, 21 - - see also Channel (Guided) wave
- see also Hypocenter Lithosphere, 3
Foreshock, 6 Love wave, 14, 15, 41, 42, 44, 50, 7 1 , 24, 25,
Free period, 60, 62 31
- vibration, 60 - - see also LQ
Frequency characteristic, 59-61 Low-velocity layer (LVL), 10, 25
see also Magnification or Response LQ, 14, 15, 19, 41-44, 48, 7 1 , 22, 24, 25, 28,
characteristic 29, 31, 34, 38, 39, 45
-, natural (eigen, free), 59 - see also Love wave
-, resonant, 59 LR, 14-16, 19, 41-44, 48, 7 1 , 12, 22, 23-26,
Frictional heat, 40 28-35, 38, 39, 42, 47, 50
- see also Rayleigh wave
g, 20, 24
G, 44, 7 1 , 30, 37, 52 Magnetic tape, 65, 67
G / , 45, 7 1 , 30 Magnification, 62
G2, G3 etc., 45 - characteristic, 60, 62
Galvanometer, 57, 58, 60 - - see also Frequency or Response characteristic
Geometrical spreading, 16 -, maximum, 60
Granitic layer, 20, 22, 24 Magnitude-frequency distribution, 8
- rock, 9 Main shock, 5, 6, 66
Greenwich Mean Time, 57 - - see also Principal earthquake
GRF, 64, 65, 77, 10, 12, 25, 47, 50, 51 Mantle, 9, 10, 28, 33, 37
Ground motion, 59 -, lower, 9
- - acceleration, 55, 57, 60 -, upper, 9, 31, 34, 40
- - characteristics, 60 - wave, 60
- - displacement, 55, 56, 60 - - of Love type, 37
- - period, 60 - - of Rayleigh type, 37
- - velocity, 55-57, 60, 64 Microearthquake, 6, 66, 68, 1, 2
Grafenberg array, see GRF Microseisms (Microseismic noise), 7, 17, 18, 61
- - see also Oceanic microseisms
Helical motion, 58 Mining tremor, 8, 25
- spring, 56 Mode (overtone), 15
Herrin tables, 51, 52 - conversion, 19, 37
Hodograph, 47 -, fundamental, 42, 43
Hypocenter, 4, 36, 47 Mohoroviiic discontinuity (Moho, M), 6, 9,
- see also Focus 20-24, 54
Monochromatic signal, 19
/, 34, 35 Moon, 45, 58
Implosion, 7, 8 Multiple conversion, 11
-, see also Collapse earthquake - rupturing, 6, 49
Industrial explosion, 60
Instrumental constant, 60 National Earthquake Information Center (NEIC),
Instrument effect, 62 32, 51, 76
International Seismological Centre (ISC), 8, 32, Nevada Test Site, 21
39, 51, 69, 2
Inverse problem, 43 Oceanic microseisms, 8, 60
 173

- - 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

pPn, 24, 70 PKIKP (PKP ), 34, 36, 37, 50, 7 1 , 55

DF

pPP, 31, 70, 48, 53 - see also P"

pPS, 31, 70 PKIKS, 34
pP'P', 48 PKJKP, 35
pS, 31, 70 PKKP, 36, 37, 7 1 , 48, 53
pSKS, 53 PKP, 33-37, 47, 49-51, 70, 55
pSP, 31 - see also P'
pSS, 31 PKPPKP ( Ρ ' Ρ ' ) , 36, 71
P, 11, 19, 2 1 , 23, 25-38, 40, 41, 43, 44, 47, PKPl (PKP ), 34-36, 7 1 , 55
BC

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

-, Press-Ewing, 17, 18, 55, 58, 61, 62 Translation motion, 55

-, S-13, 62 Travel times, 47, 49-51
-, SK, 62 - -, azimuth-dependent, 49
-, SKM-3, 62 - -, regional, 49, 50
-, strain (strainmeter), 56-58 Travel-time curve (chart), 47, 48, 4
-, variable-capacitance see Electrostatic Travel-time tables, 47-52
seismometer for near events, 53, 54
-, variable-reluctance, 57 , provisional, 50
-, Wielandt-Streckeisen, 56 Trench, 3, 4
-, Willmore, 62 TRg, 27, 70
Seismoscope, 58 TSg, 27, 70
Sg, 20-25, 28, 53, 54, 66, 70, 5, 6, 10, 11, 13
Shadow zone, 32-34 U, 42
Shear (Transverse) wave, 11, 13, 20 Underground nuclear explosion, 27, 50, 69, 72,
- - see also S wave 74-76, 14, 21, 32, 36, 38, 46
SKIKP, 34 Unipolar onset, 37
SKIKS, 34 - - see also One-sided onset
SKP, 33, 71 Universal Co-ordianted Time, 57
SKS, 33, 36, 37, 47, 49, 70, 40, 48, 51, 53 Velocity, 11
SmKS, 37, 71 -, compressional-wave, 11
SmS, 20, 23, 70, 4 - depth function, 4
Sn, 21, 23-25, 28, 29, 53, 54, 70, 10, 13 - dispersion, 15
SOFAR channel, 26, 27, 7, 14 see also Dispersion
Sonic boom, 69, 20 -, group, 42
Source mechanism, 16 -, phase, 42
- parameters, 72 -, shear-wave, 11
- radiation characteristics, 16 Vesuvius Observatory, 7
SP, 29, 70, 48, 53 Volcanic tremor, 7, 75, 76, 19
Space orbiter Columbia, 20
SPP, 53 W, 44
SS, 29-31, 47, 70, 27, 33, 34, 37, 39, 40, 42, Wadati-Benioff zone, 3, 4
43, 45, 50, 51 see also Subduction zone
SSS, 29, 70, 34, 50 Wave, 13
SSSS, 34 -, acoustic water, 7
Subduction zone, 3, 4, 6, 40, 13, 34 -, bounded, 13
- -, see also Wadati-Benioff zone -, channel, 22, 24, 41
Surface wave, 13, 14, 27, 32, 38, 40-44, 47, 50, -, core, 20, 33
60, 12, 28, 32, 34, 39 -, crustal, 20, 21, 50
- -, continental-path, 22, 28 -, direct (Elementary or Main), 29, 32, 33
- - dispersion, 22 -, free, 13
- -, fundamental mode, 15, 12, 39, 44, 50 -, guided, 13, 24, 25
- -, higher mode, 15, 12, 35, 36, 39 -, head, 21
- -, long-period, 44 -, mantle, 44
- -, oceanic-path, 28, 34 -, monochromatic, 27, 42, 19, 49
S (Secondary) wave, 14, 16-20 -, seismic, 13
- - see also Shear wave -, sound, 26, 27
Sweeping period, 50 -, tertiary (7), 26
Swinging arcs method, 48 Wiederkehrwelle, 44
World Wide Standardized Seismograph Network
T, 26-28, 70, 7, 14, 49 (WWSSN), 61, 62, 64, 65, 14, 28
TKOBS, 7, 11 WWV, 57
TK02, 7, 11
TPg, 27, 70 400 km discontinuity, 10
Transducer (Sensor, Detector), 55, 56, 58 650 km discontinuity, 10
Transform fault, 3, 4, 6
Transition zone (layer), 10, l i
 177
GEOGRAPHICAL INDEX

Numbers in boldface relate to Plate entries.

Adirondack, 77 French Polynesia, 14

Aegean arc, 38 FRG, 18, 57, 64, 65, 77, 10, 12, 25, 47, 50, 51
Afghanistan, 73, 74, 37, 43 Friuli, 12
Alaska, 75, 77, 30
Albania, 72, 27 Germany, 73, 75
Albuquerque, 77 Golden, 32
Alpide belt, 5 Great Britain, 57
Andreanof Islands, 73, 34 Greece, 17, 73, 26
Arenal volcano, 76, 17 Greenland Sea, 72
Asama volcano, 73-75, 15-17, 19 Guatemala, 76, 77, 15, 18
Atlantic Ocean, 72, 31 Guerrero, 75
Austria, 73 Goteborg, 77, 5
Azores, 5
Himalayas, 38
Bayern, 18, 64, 65, 77, 10, 12, 25, 47, 50, 51 Hindu Kush, 38
Berkeley, 77, 20, 28-30, 34, 42 Honshu, 73-75, 77, 78, 13, 48
Berkshire, 8 Hungary, 75, 22
Black Hills, 77
Bohemia, 2 Iceland, 72, 24
Brazil, 75, 41 India, 72, 44
California, 77, 20, 2 1 , 28-30, 34, 42, 46, 48, 49, Italy, 73, 12, 23
53 Izu peninsula, 74, 3

Canada, 77 Jamestown, 77, 2 1 , 46, 48

Carpathians, 38 Japan, 66, 67, 73, 74, 6, 48
Celebes, 72, 54
Celebes Sea, 72, 52 Kamata, 77, 3
Central America, 38 Kamchatka, 5, 72
Cerro Adams, 77 KaSperske Hory, 77, 37
Charleston, 5 Kazakh, 74, 32
Chiang Mai, 37, 77 Kermadec Islands, 76, 49
Chile, 45 Kiruna, 77, 27, 35, 45, 52
China, 72, 76, 33, 38, 44
Circum-Pacific belt, 5, 40 Langenthal, 4
Clyde, 77, 8 Linthal-Limmern, 77
College Outpost, 77 Longmire, 77
Colorado, 32 Marianas, 38
Cook Islands, 77 Matsushiro, 66, 67, 75, 77, 1, 6, 9, 13
Costa Rica, 74-77, 16, 17, 19, 50 Meloy, 6
Crete, 72, 26 Mexico, 66, 75, 28, 29
Cumberland Plateau, 77 Michoacan, 66, 75, 28, 29
Czechoslovakia, 77, 2, 37 Mid-Atlantic Ridge, 5
Mid-Indian Rise, 74, 55
England, 8, 32 Mina, 77, 20
Mindanao, 74
Fiji Islands, 37, 76 Molucca, 72, 54
Fiji-Tonga area, 39 Mount Hamilton, 77, 53
Fiordland region, 8 Mount St. Helens, 47
Flores Sea, 39
178

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,