of 4
URBAN
EARTHQUAKES
1995
Presidential
Address
Seismological
Society
of
America
Thomas
H.
Heaton
U.S.
Geological
Survey
525
S.
Wilson
Ave
Pasadena,
CA
91106
Although
I
have
met
many
Californians
who
are
terrified
of
earthquakes,
there
have
actually
been
relatively
few
fatalities
in
the
past
75
years.
We
have
relatively
strict
building
codes
in
the
United
States
and
this
has
undoubtedly
saved
numerous
lives.
However,
what
is
the
vision
for
our
future?
Will
we
have
man-
ageable
earthquakes
or
will
we
have
some
true
catastrophes?
One
extremely
optimistic
view
is
from
a
book
by
Robert
Hill
entitled
"Southern
California
Geology
and
Los
Angeles
Earth-
quakes"
published
by
the
Southern
California
Academy
of
Sci-
ences
in
1928.
The
following
quote
is
from
the
book
cover.
"This
book
completely
refutes
the
prediction
of
Professor
Bailey
Willis
that
Los
Angeles
is
about
to
be
destroyed
by
earthquakes.
It
proves
that
this
area
is
not
only
free
from
the
probability
of
severe
seismic
disturbances,
but
has
the
least
to
fear
from
Acts
of
God
of
any
city
under
the
American
flag."
Given
the
occurrences
of
the
1933
Long
Beach
earthquake,
the
1971
San
Fernando
earthquake,
and
the
1994
Northridge
earth-
quake,
this
claim
looks
pretty
silly.
Yet,
we
did
make
it
through
these
earthquakes
and
Los
Angeles
is
still
standing.
In
the
after-
math
of
the
Northridge
earthquake,
we
have
not
seen
a
massive
overhaul
of
our
building
codes.
Are
we
doing
most
of
the
right
things,
or
will
future
earthquakes
send
us
back
to
the
drawing
board
to
completely
rethink
things?
What
earthquakes
are
in
our
future?
What
if
we
were
simply
to
replay
historic
events
but
with
modern
populations?
We
are
certain
to
replay
the
1906
San
Francisco
earthquake.
Or
what
about
the
1700
Cascadia
earthquake
(Satake
et
al.,
1995)?
Was
it
really
a
magnitude
9
plus?
What
about
three
large
earthquakes
in
the
New
Madrid
region
in
1811
and
1812?
What
about
the
1952
Kern
County
earthquake?
Could
an
earthquake
like
this
occur
beneath
Los
Angeles?
Many
tens
of
millions
now
live
in
areas
that
were
sparsely
populated
when
they
were
last
visited
by
large
historic
earthquakes.
Do
we
know
about
what
will
really
happen
when
a
large
earth-
quake
strikes
one
of
our
urban
areas?
There
are
currently
about
30
strong
motion
records
taken
at
distances
less
than
5
km
from
earthquakes
larger
than
M
6.5,
and
they
confirm
numerous
eyewitness
reports
that
near-source
shaking
can
be
very
violent.
Peak
ground
acceleration
of
80%
g
and
peak
ground
velocity
of
100
cm/sec
are
median
values
for
these
30
records
(Heaton
et
al.,
1995).
These
values
are
considerably
larger
than
is
antici-
pated
in
our
existing
building
codes.
Although
building
codes
are
our
first
line
of
defense
against
earthquakes,
I
must
confess
that
I
don't
feel
comfortable
talking
about
them
because
I
don't
know
very
much
about
them.
Unfor-
tunately,
I
suspect
that
most
seismologists
are
in
the
same
situa-
tion.
However,
after
working
with
my
engineering
colleagues,
one
thing
has
been
made
clear
to
me.
That
is,
building
codes
are
not
based
on
an
understanding
of
ground
motions
and
the
corresponding
response
of
buildings.
Rather
they
are
largely
empirically
based.
Following
damaging
earthquakes,
engineers
study
the
performance
of
various
designs
and
modify
the
code
if
certain
designs
are
found
to
perform
poorly.
While
this
is
a
very
practical
approach,
it
may
not
predict
the
performance
of
struc-
tures
in
very
large
earthquakes,
since
the
largest
earthquake
in
a
modern
urban
environment
is
the
M
6.7
Northridge
earthquake
in
the
US,
and
now
the
M
6.9
Kobe
earthquake
in
Japan.
Even
in
the
M
6.7
Northridge
earthquake,
there
were
disturbing
indications
of
poor
performance
of
modern
structures.
Most
steel-frame
buildings
located
within
10
km
of
the
rupture
sur-
face
have
shown
serious
signs
of
distress
(Bertero
et
al.,
1994).
That
is,
cracks
have
been
discovered
in
the
joints.
These
cracks
were
a
great
surprise.
Yet
the
steel
frame
buildings
by
and
large
were
not
located
in
the
region
of
maximum
long-period
shak-
ing.
The
maximum
ground
velocities
of
over
120
cm/sec
were
in
the
northern
part
of
the
San
Fernando
valley
and
especially
in
the
Santa
Susanna
mountains
where
there
were
very
few
steel
frame
buildings
(Heaton
and
Wald,
1994).
In
fact,
a
ground
velocity
of
175
cm/sec
was
recorded
in
this
region.
The
maxi-
mum
velocities
in
regions
of
steel
frame
buildings
were
prob-
ably
less
than
50
cm/sec.
Of
course,
the
dense
downtown
part
of
Los
Angeles
only
experienced
velocities
in
the
range
of
30
cm/
sec.
I
suspect
that
we
were
quite
fortunate
that
the
Northridge
earthquake
fault
was
dipping
to
the
south.
We
can
only
guess
what
might
have
happened
if
the
same
earthquake
had
directed
its
1-second
energy
towards
downtown
Los
Angeles
on
a
north-
dipping
fault.
Seismological
Research
Letters
Volume
66,
Number
5,
September-October
1995
37
I
suspect
that
high
velocities
from
rupture
directivity
are
part
of
the
reason
that
the
Kobe
earthquake
was
so
devastating.
That
is,
a
large
concentration
of
buildings
was
subjected
to
relatively
high
ground
velocities
(Wald,
1995).
There
are
undoubtedly
many
other
factors
contributing
to
the
Kobe
disaster.
To
be
sure,
the
style
of
traditional
Japanese
wood-frame
house
construction
is
clearly
inferior
to
that
in
the
US.
Nevertheless,
in
a
country
known
for
its
attention
to
detail,
many
modern
buildings
were
destroyed
by
a
M
6.9
earthquake.
As
disturbing
as
the
Northridge
and
Kobe
earthquakes
are,
they
still
do
not
answer
the
question
of
what
might
happen
in
a
much
larger
urban
earthquake.
I
recently
visited
a
trench
opened
by
Charlie
Rubin
in
a
scarp
along
the
base
of
the
San
Gabriel
mountains
about
2
km
north
of
where
I
live
and
work.
It
showed
several
Holocene
offsets
of
about
2
to
3
meters.
I
couldn't
help
but
wonder
what
kind
of
ground
shaking
happened
in
these
prehistoric
earthquakes.
How
would
our
buildings
survive
in
this
shaking?
I
was
reading
about
Oldham's
description
of
the
1897
Indian
earthquake.
Of
course
little
definite
is
known,
except
that
very
violent
shaking
occurred
over
a
region
of
about
250
km
by
150
km.
Richter's
(1958)
book
says
that,
"not
merely
did
eyewit-
nesses
report
seeing
pebbles
bouncing
on
the
ground
'like
peas
on
a
drumhead,'
but
numerous
instances
were
observed,
photo-
graphed,
and
figured
in
detail,
of
posts
shot
out
of
their
holes
and
of
boulders
lifted
out
of
the
ground
without
cutting
the
edges
of
their
former
seats."
What
happened
here?
Could
it
happen
again?
Could
it
happen
anywhere?
At
present,
we
have
only
two
near-source
ground
motion
records
from
earthquakes
larger
than
M
7.
The
Lucerne
record
taken
in
the
middle
of
the
Mojave
desert
from
the
M
7.2
Landers
earthquake
(Iwan
and
Chen,
1994)
and
the
Tabas
record
from
the
1978
M
7.4
Tabas
earthquake
(Shoja-Taheri
and
Anderson,
1988).
Both
of
these
are
impressive
records,
the
peak
velocities
are
155
cm/sec
and
125
cm/sec,
respectively.
The
peak
ground-
motion
displacement
in
the
Lucerne
record
is
over
2
meters
and
it
is
probably
comparable
in
the
Tabas
record.
While
these
are
very
impressive
records,
it
is
unlikely
that
they
represent
the
strongest
shaking
that
occurred
in
either
of
these
earthquakes.
Furthermore,
these
are
not
particularly
large
earthquakes
when
compared
with
others
in
our
past.
In
1857,
the
slip
on
the
San
Andreas
may
have
been
about
10
meters
in
the
Carrizo
plain
(Sieh,
1978;
Grant
and
Donnellan,
1994),
or
in
1906
it
may
have
been
about
7
meters
north
of
San
Francisco
(Thatcher,
1975).
Ground
motions
were
large
enough
to
snap
trees
near
their
base
in
both
of
these
great
California
earthquakes
(see
Wood,
1955,
for
the
1857
earthquake;
see
Lawson
et
al.,
1908,
for
the
1906
earthquake).
How
do
you
snap
a
tree?
While
we
may
not
have
a
very
clear
idea
of
what
can
happen
in
close
to
a
very
large
earthquake,
we
can
come
to
some
tentative
conclusions.
The
ground
accelerations
would
be
expected
to
be
at
least
as
large
as
for
smaller
earthquakes,
but
they
are
likely
to
last
longer
and
to
occur
over
a
larger
region.
Since
we
already
have
seen
that
the
median
peak
acceleration
from
the
near-
source
region
of
smaller
earthquakes
is
about
80%
g,
we
should
not
be
surprised
by
violent
shaking
in
future
large
earthquakes.
The
most
important
difference
between
a
smaller
and
larger
earthquake
is
that
the
larger
earthquake
will
have
much
larger
slips
than
the
smaller
one.
This
means
that
the
peak
ground-
motion
displacements
can
be
much
larger
than
has
been
seen
in
the
past.
I
am
not
aware
of
any
modern
city
that
has
experienced
a
ground-motion
displacement
approaching
a
meter--perhaps
Tangshan,
China,
did
in
1976,
but
they
had
very
poor
construc-
tion
(a
large
percentage
of
the
population
was
killed).
San
Francisco
may
have
had
large
long-period
motions
in
1906,
but
the
indications
are
that
the
epicenter
was
in
the
city
(Boore,
1977)
and
thus
the
city
may
have
been
spared
from
a
severe
displacement
pulse
caused
by
directivity.
Even
so,
the
damage
from
the
San
Francisco
earthquake
was
severe.
The
1923
Kanto
earthquake
beneath
Tokyo
probably
also
produced
large
ground-motion
displacements
in
a
city,
but
again
it
is
hard
to
learn
the
appropriate
lessons
from
this
earthquake,
except
that
it
devastated
the
largest
city
in
Japan.
What
will
happen
when
these
earthquakes
repeat?
If
the
average
slip
on
a
fault
beneath
a
city
is
several
meters,
then
we
can
expect
to
see
ground-motion
displacements
of
meters.
Furthermore,
these
ground-motion
displacements
will
probably
happen
relatively
quickly;
remember
that
the
expected
veloci-
ties
probably
exceed
1
m/sec.
In
the
1983
Borah
Peak
earth-
quake,
eye
witnesses
reported
that
a
1.5
m
scarp
formed
in
about
a
second
(Wallace,
R.E.,
1984).
Similarly
short
times
were
reported
for
the
surface
faulting
from
the
Luzon,
Philippines,
earthquake
in
1990
(Yomogida
and
Nakata,
1994).
These
large
ground-motion
displacements
could
be
especially
important
for
flexible
structures,
such
as
moment
resisting
frame
buildings
and
base
isolated
buildings.
However,
there
is
virtu-
ally
no
past
experience
with
flexible
structures
that
have
experi-
enced
very
large
ground-motion
displacements.
As
it
turns
out,
the
strain
(or
drift)
in
a
building
due
to
a
wave
traveling
in
the
building
is
approximately
the
ground
velocity
divided
by
the
wave
velocity
in
the
building
(for
example,
see
Hall
et
al.,
1995).
Since
the
waves
can
interfere
with
each
other
due
to
reflections,
and
because
buildings
are
nonlinear
for
large
motions,
the
maximum
strain
in
real
buildings
could
be
even
larger.
Recent
modeling
of
the
response
of
moment
resisting
frame
buildings
to
near-source
ground
motions
indicates
that
drifts
in
excess
of
5%
could
easily
develop
in
the
lower
floors
(Hall
et
al.,
1995).
This
is
would
apply
for
both
steel
and
concrete
frame
buildings.
Although
it
seems
pretty
clear
that
buildings
would
be
heavily
damaged
by
these
large
strains,
it
is
unknown
if
they
would
collapse.
It
is
clear
that
they
will
be
taken
far
into
the
inelastic
range
of
deformation.
If
the
frame
is
not
ductile,
as
is
the
case
for
most
concrete
frame
buildings
built
before
the
mid
1970's
and
perhaps
welded
steel
frame
buildings,
there
is
an
38
Seismological
Research
Letters
Volume
66,
Number
5,
September-October
1995
even
greater
concern
that
the
buildings
would
collapse.
Further-
more,
these
large
ground
motions
have
the
capacity
to
over-
whelm
the
displacement
limits
for
existing
base-isolated
build-
ings.
In
fact,
ground
motions
already
recorded
in
the
Northridge
earthquake
are
large
enough
to
cause
violent
impacts
of
existing
base-isolated
buildings
with
their
concrete
retaining
walls
(Hall
et
al.,
1995).
So
let's
recap
the
situation.
The
recent
experiences
of
the
Northridge
earthquake
and
the
Kobe
earthquake
have
already
shown
us
that
urban
earthquakes
can
be
very
damaging
and
very
dangerous.
Furthermore,
we
can
anticipate
that
even
larger
urban
earthquakes
are
certainly
lurking
somewhere
in
our
fu-
ture.
We
have
no
direct
experience
with
such
earthquakes,
but
we
can
anticipate
large
ground
motions
having
the
capability
to
strain
buildings
far
more
than
has
been
seen
in
the
past.
These
earthquakes
also
have
the
potential
for
completely
disrupting
our
transportation
systems,
water
and
power
delivery,
sewage
removal,
and
many
other
aspects
of
our
lifelines.
Furthermore,
in
a
very
large
earthquake,
this
could
happen
over
a
very
broad
geographic
region.
Frankly,
when
I
think
of
these
things,
it
frightens
me.
Perhaps
the
one
bright
spot
in
this
picture
is
that
large
earthquakes
are
rare
and
our
lives
are
short,
so
there
is
still
a
pretty
good
chance
that
we'll
already
be
dead
by
the
time
that
a
very
large
earth-
quake
visits
one
of
our
major
cities.
Although
it
seems
that
counting
on
the
brevity
of
our
life
spans
is
the
basis
for
much
of
our
public
policy
making,
this
is
clearly
shortsighted
and
could
backfire
in
a
few
tens
of
seconds.
most
of
these
buildings
have
not
been
told
of
the
problems.
The
owners
of
the
buildings
could
suffer
catastrophic
financial
loss
if
occupants
are
told
that
the
safety
of
their
buildings
is
uncertain.
Practicing
engineers
are
paid
by
building
owners
and
it
is
diffi-
cult
for
them
to
publicly
discuss
their
concerns
about
these
buildings.
Seismologists
don't
say
much
about
the
situation
because
they
know
so
little
about
buildings.
Although
there
are
many
interests
to
protect,
who
protects
the
interests
of
the
occupants?
As
many
of
you
know,
I
have
expressed
similar
concerns
about
the
potential
dangers
from
large
ground
motions
before
(e.g.
Heaton,
1991).
Frankly,
some
earthquake
professionals
have
scolded
me
for
expressing
these
views.
They
argue
that,
while
recent
earthquakes
showed
that
there
are
some
problems,
par-
ticularly
with
older
buildings,
they
also
showed
that
modern
buildings
have
the
capacity
to
withstand
the
strongest
shaking
anticipated.
These
people
have
also
told
me
that
to
voice
my
concerns
publicly
is
irresponsible.
Why
focus
on
an
improbable
doomsday
scenario
that
scares
the
public
and
could
have
a
negative
economic
impact?
While
I'll
concede
that
doing
our
science
in
a
public
arena
is
less
than
satisfying,
I
think
that
it
is
the
best
way
to
focus
the
necessary
attention
to
understand
and
mitigate
our
vulnerability
to
future
disasters.
How
can
we
expect
to
obtain
the
resources
to
solve
these
problems
if
we
don't
acknowledge
that
there
is
a
problem?
If
we
do
not
clearly
and
openly
discuss
these
critical
issues,
then
there
are
many
compel-
ling
reasons
for
people
to
assume
that
the
status
quo
is
ad-
equate.
Unfortunately,
learning
that
the
status
quo
is
inadequate
could
come
at
a
very
high
price.
So
what
should
be
done
about
the
problem
of
urban
earth-
quakes?
Given
the
tens
of
millions
of
people
at
risk
and
the
prospect
of
a
trillion-dollar
catastrophe,
it
would
seem
logical
that
we
should
invest
heavily
in
understanding
how
to
protect
ourselves.
Yet
we
see
an
earthquake
community
that
is
strug-
gling.
I
won't
go
into
the
details
here--you
know
the
situation.
Furthermore,
if
the
funding
problem
seems
bad
to
earth
scien-
tists,
in
some
ways
it
is
even
worse
for
earthquake
engineering.
Critical
testing
has
not
been
done
and
attempts
at
mathematical
models
have
been
relatively
modest.
Given
the
extreme
cost
of
retrofitting
buildings,
it
seems
totally
illogical
to
continue
constructing
buildings
whose
response
to
future
earthquakes
is
unknown.
While
I
suspect
that
it
is
feasible
to
economically
construct
buildings
that
would
survive
very
strong
shaking,
I
doubt
that
we
have
a
clear
idea
of
exactly
how
to
do
it.
Only
through
research
by
both
seismologists
and
engi-
neers
can
we
gain
the
necessary
knowledge
to
best
design
buildings
to
survive
such
earthquakes.
I
think
that
much
of
the
problem
is
that
the
general
populace
has
no
knowledge
of
the
issues
that
I
have
raised.
Furthermore,
they
are
not
being
told.
Consider
the
problem
of
the
hundreds
of
steel-frame
buildings
with
welds
that
were
cracked
by
the
Northridge
earthquake.
As
nearly
as
I
can
tell,
the
occupants
of
I
feel
that
we
have
been
evading
key
issues
about
the
potential
impact
of
large
earthquakes
on
cities.
The
social
and
economic
pressures
associated
with
these
issues
are
so
intense
that
many
earthquake
researchers
avoid
direct
discussions
of
our
ability
to
predict
the
consequences
of
large
urban
earthquakes.
As
serious
as
earthquakes
like
Northridge
and
Kobe
were,
I
suspect
that
they
will
not
receive
much
attention
in
future
history
books.
These
pages
are
reserved
for
the
much
larger
earthquakes
that
will
inevitably
strike
heavily
urbanized
areas.
I
feel
very
frus-
trated
knowing
that
we
have
been
given
warning
signs
of
things
that
may
go
very
wrong
in
the
future,
but
we
lack
the
necessary
knowledge
to
take
the
steps
to
protect
ourselves.
If
most
of
the
public
knew
how
much
could
be
done,
but
that
isn't
being
done,
I
think
that
they
would
be
frustrated
too.
E~
REFERENCES
Bertero,
V.
V.,
J.
C.
Anderson,
and
H.
Krawinkler,
1994.
Performance
of
steel
building
structures
during
the
Northridge
earthquake,
UCB/
EERC
94/09,
Earthquake
Engineering
Research
Center,
University
of
California,
Berkeley.
Boore,
D.
M.,
1977.
Strong
motion
recordings
of
the
California
earthquake
of
April
16,
1906,
Bull.
Setsm.
Soc.
Am.
67,
561-576.
Grant,
L.
B.,
and
A.
Donnellan,
1994.
1855
and
1991
surveys
of
the
San
Andreas
fault:
implications
for
fault
mechanics,
Bull.
Seism.
Soc.
Am.
84,
241-246.
Hall,
J.
F.,
T.
H.
Heaton,
M.
Hailing,
and
D.J.
Wald,
1995.
Near-source
Seismological
Research
Letters
Volume
66,
Number
5,
September-October
1995
39
ground
motion
and
its
effects
on
flexible
buildings,
submitted
to
Earthquake
Spectra.
Heaton,
T.
H.,
1991.
Near-field
ground
motions
in
large
earthquakes
and
base-isolated
structures,
Seism.
Res.
Lett.
62,
no.
1,
49.
Heaton,
T.
H.,
and
D.J.
Wald,
1994.
Strong
ground
motions
from
the
Northridge
earthquake:
were
they
anomalous
or
a
warning
of
things
to
come?
EOS,
Trans.
Am
Geophys.
U.
75,
no.
44,
175.
Heaton,
T.
H.,
J.
F.
Hall,
D.J.
Wald,
and
M.
Hailing,
1995.
Response
of
high-rise
and
base-isolated
buildings
to
a
hypothetical
M
w
7.0
blind
thrust
earthquake,
Science
267,
206-211.
Hill,
R.
T.,
1928.
Southern
California
Geology
and
Los
Angeles
Earth-
quakes,
Southern
California
Academy
of
Sciences,
Los
Angeles,
CA,
232
p.
Iwan,
W.
D.,
and
X.
D.
Chen,
1994.
Important
near-field
ground
motion
data
from
the
Landers
earthquake,
Proceedings
of
the
1~
h
European
Conference
on
Earthquake
Engineering,
Vienna.
Lawson,
A.
C.,
and
others,
1908.
Report
of
the
State
Earthquake
Investiga-
tion
Commission,
Carnegie
Institution
of
Washington,
Publication
No.
87.
Richter,
C.
F.,
1958.
Elementary
Seismology,
W.H.
Freeman
and
Company,
Inc.,
San
Francisco.
Satake,
K.,
K.
Shimazaki,
and
Y.
Tsuji,
1995.
A
possible
Cascadia
earth-
quake
of
January
26,
1700
as
inferred
from
tsunami
records
in
Japan,
seism.
Res.
Lett.
66,
no.
2,
44.
Sieh,
K.
E.,
1978.
Slip
along
the
San
Andreas
fault
associated
with
the
great
1857
earthquake,
Bull.
Seism.
Soc.
Am.
68,
1421-1448.
Shoja-Taheri,
J.,
andJ.
G.
Anderson,
1988.
The
1978,
Tabas,
Iran,
earth-
quake:
an
interpretation
of
the
strong
motion
records,
Bull.
Seism.
Soc.
Am.
78,
142-171.
Thatcher,
W.,
1975.
Strain
accumulation
and
release
mechanism
of
the
1906
San
Francisco
earthquake,
J.
Geophys.
Res.
80,
4862-4872.
Wald,
D.J.,
1995.
A
preliminary
dislocation
model
for
the
1995
Kobe
(Hyogo-ken
Nanbu),
Japan,
earthquake
determined
from
strong
motion
and
teleseismic
waveforms,
Seism.
Res.
Lett.
66,
no.
4,
22-28.
Wallace,
R.
E.,
1984.
Eyewitness
account
of
surface
faulting
during
the
earthquake
of
28
October,
1983
Borah
Peak,
Idaho,
Bull.
Setsm.
Soc.
Am.
74,
1091-1094.
Wood,
H.
O.,
1955.
The
1857
earthquake
in
California,
Bull.
Seism.
Soc.
Am.
45,
47-67.
Yomogida,
K.,
and
T.
Nakata,
1994.
Large
slip
velocity
of
the
surface
rupture
associated
with
the
1990
Luzon
earthquake,
Geophys.
Res.
Let.
21,
1799-1802.
40
Seismological
Research
Letters
Volume
66,
Number
5,
September-October
1995