Mechanism of the 1975 Kalapana, Hawaii, earthquake inferred from tsunami data
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. B6, PAGES 13,153-13,167, JUNE 10, 1999
Mechanism of the 1975 Kalapana, Hawaii, earthquake inferred
from tsunami data
Kuo-Fong Ma
Institute of Geophysics, National Central University, Chung-Li, Taiwan
Hiroo Kanamori
Seismological Laboratory, California Institute of Technology, Pasadena
Kenji Satake
Seismotectonics Section, Geological Survey of Japan, Tsukuba, Japan
Abstract. We investigated the source mechanism of the 1975 Kalapana, Hawaii, earthquake
(Ms= 7.2) by modeling the tsunamis observed at three tide-gauge stations, Hilo, Kahului, and
Honolulu. We computed synthetic tsunamis for various fault models. The arrival times and the
amplitudes of the synthetic tsunamis computed for Ando's fault model (fault length= 40 km, fault
width= 20 km, strike= N70°E, dip= 20°SE, rake= -90°, fault depth= 10 km, and slip= 5.6 m)
are - 10 min earlier and 5 times smaller than those of the observed, respectively. We tested fault
models with different dip angles and depths. Models with a northwest dip direction yield larger
tsunami amplitudes than those with a southeast dip direction. Models with shallower fault depths
produce later first arrivals than deeper models. We also considered the effects of the Hilina fault
system, but its contribution to tsunami excitation is insignificant. This suggests that another
mechanism is required to explain the tsunamis. One plausible model is a propagating slump
model w ith a l m subsidence along the coast and a 1 m uplift offshore. This model can explain
the arrival times and the amplitudes of the observed tsunamis satisfactorily. An alternative model
is a wider fault model that dips l 0 °NW, with its fault plane extending 25 km offshore, well beyond
the aftershock area of the Kalapana earthquake. These two models produce a similar uplift
pattern offshore and, kinematically, are indistinguishable as far as tsunami excitation is concerned.
The total volume of displaced water is estimated to be ~2.5 km 3 . From the comparison of slump
model and the single-force model suggested earlier from seismological data we prefer a
combination of faulting and large-scale slumping on the south flank of Kilauea volcano as the most
appropriate model for the 1975 Kalapana earthquake. Two basic mechanisms have been
presented for explaining the deformation of the south flank of Kilauea: ( l) pressure and density
variation along the rift zone caused by magma injection and (2) gravitational instability due to the
steep topography of the south flank of Kilauea. In either mechanism, large displacements on the
south flank are involved that are responsible for the observed large tsu namis.
1. Introduction Several studies have been made to determine the mechanism
of the 1975 Kalapana earthquake using seismic, aftershock.
The 1975 Kalapana, Hawaii, earthquake (M s = 7.2) is the geodetic. and tsunami data. From the analysis of seismic waves,
largest Hawaiian earthquake instrumentally recorded. It tsunami_ and crustal deformation data, Ando [1979] obtained a·
occurred at 04 47:30 Hawaiian Standard Time (14 47:30 gmt) on normal-fault mechanism for this earthquake with strike= N70°E:
November 29, 1975 [Tilling et al.. 1976], and the location given dip= 20°SE; fault length = 40 km: width = 20-30 km: depth = I 0
by Hawaiian Volcano Observatory tHVO) is at I 9°20'N, km: slip = 5.5-3.7 m. Furumoto and Kovach {1979], using
J 55°02' W and a depth of 5-7 km (Figure 1a). This earthquake teleseismic and local seismic data, obtained an overthrust
affected most of the south flank of Kilauea volcano between the mechanism with a dip of 4°NW and a strike of N64°E . These
southwest rift zone and the east rift zone (Figure I b) and was solutions are shown in Figure I c. Crosson and Endo [ 1982]
accompanied by large tsunamis that caused significant damage. made extensive studies of south flank earthquakes, suggested a
The tsunamis were observed at several locations along the coast. decoupling layer that is the old oceanic sediment layer at the base
A severe tsunami reached a maximum height of 14.6 m at Halape of the volcanic pile on the south flank of Kilauea, and proposed a
beach (Figure lb), where two campers were killed. Similar mechanical model for south flank deformation that involves slip
large earthquakes accompanied by tsunamis previously occurred motion occurring on top of the old oceanic crust. These
in this region of the island in 1868 and 1823. seismological studies suggested that the Kalapana earthquake
occurred on a nearly horizontal fault plane at a depth of 7-9 km.
C opyright 1999 by the American Geophysical Union.
although the dip angle of this plane differed among different
Paper number 1999JB900073 . investigators. Wyss and Kovach [1988] argued for a nearly
0148-0227 /99!1 999 JB900073 $09. 00 horizontal thrust faulting mechanism with secondary normal
13,15313,154 MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE
a Bathymetry b c
23 0 0
22
21 Ando [1979]
20
19
f!:J
Furumoto and
Kovach [1979]
18
-159 -158 -157 -156 -155 -154
Longitude (deg.)
Figure 1. (a) The epicenter (star) and the fault plane of the 1975 Kalapana earthquake and three tide-g11uge
stations (triangle). The bathymetry around the Hawaiian Islands is shown by contour lines with intervals of 1000
m. (b) Kilauea (K), east rift zone (ERZ), southwest ritl zone (SWRZ), and the Hilina fault system. (c) The
focal mechanism (equal-area projection of the lower focal hemisphere) of .the 1975 Kalapana earthquake
determined by Ando [I 979] and Furumoto and Kovach [ 1979]. The shaded and open areas represent compression
and dilatation. respectively.
faulting that may have involved gravitational slumping. More earthquake was reported by Lipman et a/. [ 1985] (Figure 2).
recently, Bryan and Johnson [1991] analyzed the mechanisms of The maximum displacement occurred at Halape with subsidence
earthquakes on the island of Hawaii from l 986 to 1989 and of 3.5 m. The amounr of subsidence along the south flank of
suggested that Kilauea's south flank is mobile and moving Kilauea decreases abruptly to the west of Halape and more
seaward. gradually to the east. The horizontal extensions observed by
The vertical displacement associated with the 1975 Kalapana Lipman el al. (1985] show a steadily increasing trend seaward
155°30' 155°15' 155. 0
l9°45'~~~~~~~~~~~~~~~~~'TT~~~~~~~ .....
I'"
/
1"'
I
/
I
I
19 °30 • I
/
,..-/
./'.~---
1 9° 15' 15krn
Figure 2. The vertical displacement associated with the 1975 Kalapana earthquake [after Lipman er al., 1985] and
the cross section line AA'. The contour lines indi cate the uplift (positive) and subsidence (negative) in meters
(simplified from Lipman et al. JI 985]). A ' A" extends the cross-section line of AA· to offshore with the same
di stance.MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE 13,155
over the south flank (Figure 3) and ground cracks along the 25 large tsunamis. According to this result. if the south flank fails
km of the Hilina fault system. The subaerial horizontal in a more catastrophic landslide, it will yield even larger tsunamis
displacement increased from I m near the summit of Kilauea to 8 than those generated by the 1975 Kalapana earthquake.
m at the coast. The maximum horizontal displacement occurred The 1975 Kalapana earthquake exhibited a somewhat peculiar
in the same area of the south flank as the maximum subsidence. radiation of long-period surface waves. On the basis of the
Considering the average subsidence and seaward displacement, radiation pattern of Love waves, Eissler and Kanamori [1987]
they estimated that the total volume of deformation is - 2 km3. proposed a near-horizontal single-force mechanism that
Swanson et al. [1976] noted that the entire south tlank of Kilauea represents slumping. The single force is oriented opposite to the
is mobile and has undergone extensions of several meters in the direction of the inferred slumping on the sotJth tlank of Kilauea
century prior to the 1975 Kalapana earthquake. More recently, volcano. However, Kawakatsu [1989) showed that the overall
Owen et al. [1995] concluded from Global Positioning System radiation patterns of fundamental and overtone surface waves can
(GPS) measurements that the rapid deformation of the south flank be explained with a fault model (i.e., double couple).
can be explained by slip on a low-angle fault beneath the south Now it is generally agreed that the 1975 Kalapana earthquake
flank combined with dilation deep within the Kilauea's rift involved large horizontal displacement on a nearly horizontal
system. decollement at a depth of 7-9 km on the south flank of Kilauea.
Swanson et al. (1976], Dvorak et al. [1986, 1994], and Dvorak However, the extent, the magnitude, and the timescale of the
[1994] proposed that the magma system beneath Kilauea displacement offshore are still uncertain because of the lack of
volcanoes must have pushed the south flank seaward. Denlinger geodetic data. It is essential to understand the nature of the
and Okubo [1995) considered the geological structure. seismicity, deformation field offshore for a better understanding of the
and deformation of the south flank of Kilauea and suggested that mechanism of deformation of the south flank of Kilauea. In this
the structure of the mobile south flank is probably associated with paper we model the tsunami data observed in Hawaii to gain
a magma system at a depth below 4 km. which pushes a wedge- further constraints on the mechanism of this earthquake. The
shaped portion of the south flank seaward on the decollement. tsunamis excited by this earthquake were wel! recorded at several
The decollement begins with the 7 to I 0 km deep band of tide-gauge stations around the Hawaiian Islands. Tsunami data
seismicity bordering the east rift zone and Koae fault system. provide useful constraints on crustal deformation in the offshore
They suggested the association of a localized offshore area and complement the geodetic data. which are limited to
topographic bench, 30-50 km from shore and 5 km below sea onshore areas.
level, with the decollement structure. They also suggested that We will show that an average vertical displacement of - 1 m
the magma pressure in the wedge-shaped portion of south flank over an offshore area of - 2800 km2 is necessary to explain the
of Kilauea is an important driving mechanism of the slump on the tsunamis. This uplift can be most easily explained by a large-
decollement. This result suggests that the slumping induced by scale seaward extension of the horizontal slip on the onshore
the magma intrusion is the most important factor for producing decollement.
155°3 0' 155°15 I 155°
19°30'
19°15 1
- 3m
Figure 3. The horizontal displacement associated with the 1975 Kalapana earthquake (simplified from Lipman et
al II 985]).13,156 MA ET AL.; MEC HANISM OF 1975 KALAPANA EARTHQUAKE
2. Tsunami 2.2. Method
Ts unamis from the 1975 Kalapana ~arthquake have been t sunami modeling has been done by many investigators. For
studied by several investigators. Hatori [1976), using a ts unami example, Aida [1969) performed numerical moqeling to study
ray-tracing method, estim11ted th~- possible tsunami-generating tsunamis caused by the 1964 N iigata and 1968 Tokachi-Oki
area to be 2200 km2 where an average uplift of 1 m occurred. earthquakes. The computed tsunami waveforms were in
Ando [1979] computed a synthetic tsunami at Hilo tide station for satisfactory agreement with those observed at tide-gauge stations.
his seismic fault model. His synthetics have amplitudes too Hwang et al. [ 1970, I 972a, b) simulated the tsunamis caused by
small -compared with the observed amplitudes. Here we the 1964 Alaskan, the 1960 Chilean, and the 1957 Andreanof
investiga~e the three tide gauge recprds observed at Hilo. Islands earthquakes . They developed a numerical model for
Hono lulu, and Kahului (Fig ure la) to determine the seafl oor generation and transoceanic propagation of tsunamis using
deformation that is responsible for the tsunami s and the poss ible hydrodynamic equations in a spherical coordinate system. Aida
mechanism of the 1975 Kalapana earthquake, [1 978) showed that most tsunami s can be approximately
explained with seismic fault mode ls. Recent studies by Satak.e
2.1. Data [19 87, 1989] showed that the slip distributi on on the faul t plane
of large submarine earthquakes can be determined by inversion of
Figure 4a shows the tsunami records at three tide gatige
tsunami waveforms. Ma et al. [199 1a, b] determined the
stations. Hilo, Honolulu. and Kahului, taken from Cox [1980].
vertical seafloor deformation associated with the 1989 Lorna
According to the gauge-time corrections determined by Cox
Prieta, California, and the 1906 San Franc isco. California,
[ 1980] by checking the clocks o n the day before and after the
earthquakes from tsunami data obtained at nearby tide-gauge
earthquake the average timing errors were O. -0.5, and 6.5 min for
stations. Recently, the tsunami modeling technique has been·
Hilo, Honolulu, and Kahului, respectively. The positi ve and
ex ten s i v~ly applied to determine the space and temporal
negati ve signs indicate the advance and delay of the tide. gauge
variations of sli p of important tsunamigenic and tsµnami
clocks. r~spectively. Applying the gauge-time corrections to the
earthquakes [Johnson and Satake. 1996; Tanioka and Satake.
tide-gauge records, we digitized the records fo r 1.5 hours starting
1996; Johnson and Satake, 1997]. A useful review of this
froin the origin time of the earthquake and dctrended them.
subject is given by Geist [ 1997).
Fig ure 4b s hows the time-corrected detrended records.
We take a Cartesian coordinate system (x. y . z) with the z axis
Although the records show some background noise. we can
oriented vertically. Then the basic eq uations of motion for
iden ti fy th e distinct upward first arrivals as indicated on Figure
tsunam i computation are g iven by
4h. The first arrivals thus identified at H ilo. Hon olulu, and
Kahului are at 23, 48, · and 49 min after the origi n time of lhe
earthquake, n;:spectively, with an error of - 2 miri. The peak to
01
-+u- + v -
ou Cit. OH r u(u 2 +v 2 )
- fv = - g - - - " - -- -'---
112
peak amplitudes are --180, 15, and 85 cm at Hilo , Honolulu, and a & 0- & ( H +D - 77)
Kahului, respecti vely. · (I)
a
Hilo
Honolulu Kahului
I
..l·i\I ~M .1 "M1'
:1 j1 I
. !J\A.;, :f .'
! l~!
; I
!I ~.i ': '
i· ~!
;I { I
I
...1!ir_
b
120 15 60
80 10 40
,......
a 40 5 2Q
~
.,
"O
.a 0 0 0.
r
;=I
-40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 6 30 60 90 0 30 60 90
Time (min) Time (min) Time (min)
F igure 4, (a) Tsunamis record ed on the tide gauges at Hilo. Honolulu, and Kahului . (b) Oetrended tsunami
records for 90 min starting from the origin time. Arrows indicate the onset of tsunami .MA ET AL.: MECHANIS M OF 1975 KALAPANA EARTHQUAKE 13,157
O\>
-+ u~+v~+
O\> O\> oH ru(u 2 + v 2 ) 11 2
Ju = - g - - - " " - --'---
few minutes, we cah assume that the water surface is uplifted in
a ax 0? 0?13,158 MA ET AL.: MECHAN ISM OF 1975 KALAPANA EARTI-IQUAKf,
20° cm. this resu lt suggests that an average of -1 m upliil o ver the
tsunami source area would be required to explain the observed
amplitudes.
4. Comparison With Geodetic Data
Si nce the o bserved tsunami cou ld not be simply explained by
Ando's fault model, we considered the vertical displacements
determ ined by geodetic surveys onshore [Lipman et al., 1985,
Figure 2] to coristrain the fault models. We compared the vertical
19° ground deform ation on land from various fault models with
d iffere nt dip ang les and fault depths to the observed geodet ic data.
Then we computed tsunam is for those fault models and compared
them with the. observed.
Figure 8 shows the displacement along the profile AA' passing
through Halape as shown in Figure 2. The first trough of the
displacement profil e to the north is near the summ it of Kilauea
vo lcano. The steep gradient near A ' is located near the Hilina
0 30 fau lt system and is probably caused by displacement on it. The
L___L_J
Klil segment A'A" in Figure 8 covers the offshore ar~ . From the
comparison of the tsunami amplitudes for Ando's fault model
-155° -154° with the observed we suggested earlier that an average
Fig ure 6. Inverse travel time isochrons. The solid. dash-doted. displacement of 1 m offshore is required to explain ihe observed
and dashed curves ·indicate the tsunam i wave fronts at every tsunamis. This average is indicated in Fig1.1re 8.
minute from 20 to 25. 45 to 5, and 40 to 50 mi n for l-l ilo. The dip angle of the fault model of the 1975 Kalapana
Honolulu, and Kahului stations, respectively. The bold curves earthquake is not very well determined. Ando's [1979] solution
indicate the onset-time isochrons. The asterisk indicates the is a normal fault dipping 20°SE. while Furumoto and Kovach's
epicenter of the earthquake. The shaded area represents the fault ( 1979] solution is a thrust fault dipp ing 4°NW (Figure I c).
plane. · Given this uncertainty, we computed the vertical crustal
deformation profi les alo ng AA" for fault models with dip angles
synthetics are too early in arrival time. as we expected. and too of 20°SE (Ando's [1 979] model). 10°SE. 0°. 4°NW (Furumoto
small in amplitude compared with the observed. Since the and Kovach's [1979] model) and I0°NW. The results are shown
record at Kahului has a large long-period background noise, the in figure 8. The fault parameters for these models. other than
beginning is somewhat ambiguous, but the arrival time of the first the dip angle. are the same as in Ando's model with a fau lt sli p of
peak of the synthetic is earlier than that of the o bserved. The 5.5 m. As the dip direction changes from SE to NW, the amount
fir st arrival times of the synthetics are - l 0 min too early. and the of subsidence onsho re decreases, but the amount of uplift
peak to peak amplitudes of the synthetics are about one fifth of offshore increases. The model dipping 20°SE yields a
the observed. Since the average u pli ft for Ando's model is - 20 maximum subsidence of - 170 cm and a maximum uplift of only
Ando's M odel
Hilo Honolulu Kahului
120 15 60
80 10 40
E' 40 5 20
~
ii> '
-0 "
a
:=l
0 0 0 ,,
0.
~ -40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
Time (min) Time (min) Time (min)
figure 7. Comparison of the synthetic tsunami s (dashed curve) computed fo r A ndo's fault model with the
observed (so lid curve).MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE 13,159
Depth=IO kin
300
200 average displacement
---- -
"',,.,.,./=-~
100 IJ))-Y,.))'f>))»"11W~"'>»">'h'>"1f£~)'f>-;_)'h)))'1>"'t>)))'»'>">""»~'))"r))))))))).')'1h)·'h'>))'t;
,......., /... ---~~
s /// / ____ :-~
.......,,
t.) ',, /. / , -
...... / / . ,
i:: 0 -=-~~ ... '/ / / ,'
(i)
~~.;:.
8 ,
// / / / , ,'13,160 MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE
Depth=10
Hilo Honolulu Kahului
120 15 60
80 10 40
40 5 20
0 0 0 10°SE
-40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
120 15 60
80 10 40
40 5 20
0 0 0
-40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
120 15 60
80 10 .
"
40
40 5 20
0 0 0 4°NW
-40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
120 15 60
80 10 40
40 5 20
a
~
o 0 0 10°NW
;.:::l
0. -40 -5 -20
~ -80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
Time (min) Time (min) Time (min)
Figure 9. Comparison of the synthetic tsunamis (dashed curves) for fault models with dip angles of 10°SE, 0°.
4°NW, and I 0°NW with the observed tsunamis (solid curves). The upper edge of the fault is at 10 km.
for understanding the origin of the observed tsunamis, we next 1.0
consider the displacement on the Hilina fault system. • x
Depth=10 km
Depth:o:3 km
The steep gradient of subsidence near the coast is probably due
to the slip on the Hilina fault system. The Hilina fault system is
characterized by south facing normal fault scarps as high as 500
-
0
~
a:
0.8MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE 13,161
a dip=10°SE b dip=0° c dip=l0°NW
300
average displacement
200 average displacement !\
F-~
~ 100
~
// '-.~-
~ 0
E
] -100 obs.
~ 10
cs -200 7
5
-300
-400
A
--
15km
A' AA A'
3
A' A A'
Bathymetry
-1500°
-3000
~
-
=';] I
- .
Figure 11. Comparison of the vertical crustal deformation for fault models with the upper edge of the fault at
depths of 10. 7, 5, and 3 km with the observed deformation (solid curve) inland and the required average uplift
(shaded line) offshore. Comparisons for fault models. with dip angles of (a)10°SE. (b)0°. and (c)l0°NW are
. shown. The bottom boxes indicate the bathymetry along the profile AA".
to the main fault model dipping 10°SE at depth of 3 km, we can (Figures 2 and 3 ). Combining these observations with the
explain the overall pattern of the profile AA' as shown in Figure asymmetry of the deformation (the gradient of subsidence along
13. This solution is nonunique since other models could be the south flank of Kilauea decreases more rapidly west of Halape
obtained by adjusting the amount of slip and the depth of the than east), they suggested that the initial earthquake triggered a
upper edge of the fault. However, the synthetic tsunami from sequence of deformation tbat migrated westward along the Hilina
the Hilina fault system model contributes insignificantly to the fault system. The ratio of horizontal to vertical displacement
overall tsunami excitation, as shown in Figure 14. associated with the 1975 Kalapana earthquake suggests a
gravitational slump. Lipman et al. [ 1985] also observed new
ground breakages in the Hilina fault system, as much as I m
5. Slump Model
along the southwestern part but minimal along the eastern part of
From the results of various studies mentioned above it seems the Hilina fault system; none was found near the epicenter.
reasonable to assume an essentially horizontal slip plane at a Widening of many cracks occurred within the Hilina fault system,
depth of - I 0 km. From the magnitude of the geodetic which indicated significant horizontal extensi on. The patterns
deformation the average slip on this plane is estimated to be -5 m. of ground breakage along the Hilina fault system and the leveling
As shown in Figures 9 and 10, this model explains only about contours showing the maximum gradients of subsidence in the
half the observed tsunami amplitudes. This discrepancy was same area also offer convincing evidence that much of the
already noted by Ando [1979}. One might conclude that earthquake-related deformation involved seaward gravitational
considering all the uncertainties in the tide-gauge data and the slumping or block sliding.
limitations in tsunami waveform computations (e.g., reflections As mentioned earlier, the beginnings of the upward motion are·
from the coast are ignored.), this discrepancy is not significant. at - 23. 48, and 49 min after the origin time at Hilo. Hono lulu,
However, although it is difficult to assess the real uncertainties in and Kahului, respectively. Hence, if slumping occurred at the
tsunami modeling, considering the many successful modeling same time (within a minute or so) as the earthquake, the slump
studies of the first-motion waveforms of tide gauge data source must be located near the point where the corresponding
[Imamura et al., 1993, 1995; Johnson et al., 1996; Johnson and isochrons (Figure 6) come close to each other. Considering the
Satake. 1997], we feel that it is worthwhile to try to explain this large subsidence along the Halape coast (Figure 2), we
discrepancy by modifying the conventional dislocation models. constructed a propagating slump model with subsidence in the
To remove the discrepancy between the observed and computed coastal area as shown in Figure 15a. We divided the large
tsunami shown in Figure 9, some additional uplift of the seatloor source area shown in Figure l 5a into 12 blocks, each block
is required. Although we cannot determine this source uniquely, having an area of 9x9 min. We assume a I m coseismic
we first invoke a slump feature on the seafloor. subsidence in the three blocks along the coast and a I m uplift in
Lipman et al. [1985] showed that the onshore extensional the remaining 9 blocks (-2500 km2) with delays shown in
ground deformation related to the 1975 earthquake and associated Figures. The amount of subsidence was determined from the
normal faults were as much as 3.5 m vertically and 8 m average subsidence along the coast. We estimated the time
horizontally. The maximum horizontal displacement occurred delays for each block of the propagating uplift from the isochrons
in the 5ame area of south flank as the maximum subsidence shown in Figure 6. Figure 16a shows the synthetic tsunamis13,162 MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE
Depth=3 km
Hilo Honolulu Kahului
120 15 60
80 10 40
40 5 20
0 0 0 10°SE
-40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
120 15 60
80 10 40
40 5 20
0 0 0 oo
-40 -5 -20
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
120 15 60
80 10 ,.
I 40
---
a •'
~ 40 5 •'•' 20
'MA ET AL.: MECHANISM O F 1975 KALAPANA EARTHQUAKE 13,163
:soo..----- ----,..--------, d ata and the l\mitations in tsunami waveform computations. the
deplh=10
d iscrepancy between lhe synthetics and observed is not
depth=? significant. Minor modifications of the wide fault model would
deplh=S improve the fit. Thus, if we extend the fault plane nearly 25 km
-E 100
depth=3 offshore, well beyond the aftershock area, we can exp la in the
observed tsunami without invok ing a sl umping offshore. In
-
other word s. the wide fau lt mode l an d the slump model are
u ki nematically simi lar as far as tsunami generation is concerned.
+J Here we prefer the slump model because it is more compatible
c with the absence of the aftershocks offshore and the deformation
Q)
E
Q)
pattern in the coastal area described by Lipman et al. [1 985 ].
(.) -100
cu 7. Comparison With Single Force Model
0.
(f)
-200 We now compare the slump model with the single force model
0 of Eissler and Kanamori [1987). · However, since th e single-
force mode l was obtained from seismic rad iation, it cannot be
-300 directly compared · with the present result. We make on ly a
15km q ualitati ve comparison in the following.
The single force is a kinematic representation of southeastward
slumping of a large block on the south flank of Ki lauea. The
northern half o f th is block is onshore and the southern half is
Fig ure 13. Comparison of the vertical deformation for a offshore LEissler and Kanamori. 1987, Figure 12). Seaward
composite fault model with the observed. The composite model sliding of the offs~ ore part of this wedge-like structure would
consists of a fau lt with a dip angle of I 0°SE and a depth of 3 km uplift the seatloor over a large area. The pattern of uplift is
and the Hilina fault. consistent with that determined from tsunami data. Since the
magnitude of the sing le force depends on the total mass of the
sliding block wh ile tsunami excitation depends only on the area
average uplift of - 80 cm over an area of - 2400 km2, which is and the amount of uplift, no d irect comparison can be made
similar to that fo r the slump model shown in Figure I 5a. Figure between the magnitude of the sing le force and tsunami excitation.
19 shows the synthetic tsunamis computed for the wide fatJlt The overall magnitude of the slump can be given by the
model with 5.5 m slip on the fau lt. The arrival times of the product MD using a centroid single force (CSF) model
synthetic tsunam is are d iffererit from those of the observed [Kawakatsu, 1989] , where M and D are the total mass of the
tsunamis by 2- 5 min , and the average amplitude is - 71 % of the slump and slid ing d istance. If we assume a time function of the
observed. Cons idering all the uncertainties in the tide-gauge fo rce to be a simple si nuso id with half-period T ,
Hilina Fault Model
Jlilo Honolulu Kahului
120 15 60
80 10 40
es, 40 5 20
cl)
'O
.a 0 0 0
;::!
0.
a -40 -5 -20
~
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
Time(min) Time (min) Time (min)
Figure 14. Comparison of the synthetic ts unamis (dashed curve) for a Hi lina fault model with the observed
tsunami s (solid curve). ·13,164 MA ET AL.: MECHANISM OF 1975 KALAPANA EARTIIQUAKE
a 20
19.5
alapana
19
4 4 4 4
0 Km 30
18.5 18.5
-156 -155.5 -155 -154.5 -i54 -156 -155.5 -155 -154.5 -154
Longitude (deg) Longitude (deg)
Figure 15. (a) Propagating slump model. Hatched boxes indicate subsidence; and open boxes indiMA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE 13,165
Eissler and Kanamori [1987), estimated D to be 37-370 m.
Since the observed subaerial horizontal displacement ·increased
from 1 m near the summit of Kilauea to 8 m at the coast (Figure
-..... . . . Subsidence 3). the above value of D suggests that the displacement continued
...........
..... to increase offshore as the slide block deteriorated into a massive
sediment slump. The amount of vertical displacement caused by
this horizontal displacement depends on the details of the slide
geometry. If the slide is a simple wedge with a triangi.ilar shape
with a slope a, the vertical displacement I-I is given by Dtaha.
Eissler and Kanamori's [1987) gravity slide model suggests a=
5°. This slope, however. results in H = 3 to 227 m, which is too
large compared with the uplift estimated from tsunami data. If
the average slope of the offshore slump decreases, H can be
Figure 17. Schematic showing a rotational slumping that causes reduced accordingly, Although large uncertainties are involved
subsidence and uplift in the determination of Jo and i: and the total volume and
geometry or the slide mass are not known well, the above
comparison appears to indicate that the magnitude of the single
I . Jrt
1 ,,s(t ) = f,, sm(-) 0 :S; t : 2r force is too large to be consistent with the observed tsunami.
r Kawakatsu [1989) concluded that neither a double-couple rior
(6) single force model can explain the data satisfactorily and
0 t > 2r suggested a combination of them. If part of the seismic
where fiJ is the peak force, the magnitude of the sluinp can be radiation is due to faulting, then the magnitude of the single-force
can be reduced and the resulting uplift can be made consistent
written as
with that estimated from tsunami data. This is consistent with
MD=pVD the argument of Wyss and Ko vach [1988) in which the 1975
Kalapana earthquake is viewed as a faulting onshore with
= ;;,r[ts(r")dr" }11 secondary normal faulting and slumping in the coastal and
offshore areas.
= J0 J:'[J~sin( 7r;"}•"Jdt
8. Conclusion
= 2/o ,2 (7) The syntheiic tsunam is computed for Ando's [1979)
7'
di slocation model are too early in arrival times and too small in
where p and V are the density and total volume of the slump and amplitudes. We considered the effects of the Hilina fault system,
i:" is the dummy parameter. Eissler and Kanamori [ 19871 in addition to the main fault, but the contribution of the Hilina
estimated Jo to be - 1x1015 N from long-period surface waves. fault system to tsunami excitation is insignificant. We tested a
Then they estimated D to be 80-2600 m for ranges of M and i: model that consists of faulting on a horizontal plane at a depth of
from 1015 to 1016 kg and 50 to 90 s, respectively. Ka wakatsu I 0 km, a Hilina fault model.and a simple slump (gravity faulting)
[1989) applied an inversion method. and using more data than model. We fouri.d that a propagatinlS slump model with I m
Mauna Loa
(/)
c::
w
..... 0
w
~
0
-'
~ 10
z
:r: Subaerial ·flows and
~ 20 ~ elastic deposits ~ Ultramatic cumulate
~ ~
LU
•
Cl Submarine pillow basalts Oceanic crust (layer 3)
\
Basaltic dikes
~ Uppermost mantle
30
40 30 20 10 0 10 20 30 40 50 60
DISTANCE FROM SHORELINE IN KILOMETERS
Figure 18. The location and geometry of the wide fault model (bold dashed line) shown on the cross section given
by Hill and Zucca [1987}. The extent of the standard fault models (width= 20 km) used in this study is shown.
by the bold lines.13,166 MA ET AL.: MECHANISM OF 1975 KALAPANA EARTHQUAKE
Hilo Honolulu Kahului
120 15 60
,
,1
80 IO •' 40
•',1•'
•', 1
40 5 •'
•'~
20
I
'
I
I
I
'' I
0 0 0 I
',
-40 -5 -20 '. '
-80 -10 -40
-120 -15 -60
0 30 60 90 0 30 60 90 0 30 60 90
Time (min) Time (min) Time (min)
Figure 19. Comparison of the synt hetic tsunamis (dashed curve) computed for the wid e fau lt model w ith t he
observed tsunamis (solid curve).
subsidence along the coast and I m uplift offshore can explain the fau lting due to forcdi1 l injection of magma, ./. Geophys. Res., 84.
arrival times and amplitudes of the tsunami s very well. An 76 16-7626. 1979.
alternative model is a wider fault mode l that dips I0°NW, with its Bryan. C. J .. and C. C. Johnson. Block tectonics of the island of Hawaii
from a focal mechanism analysis of basal slip, Bull. Seismal. Soc. Am..
fau lt p lane extend ing 25 k m offshore well bey ond the aftersh ock
81. 49 1-507. 199 1.
area of the Kalapana earthquake. These two mode ls produce a Cox. D. C.. Source of the tsunami associated with the Kalapana (Hawaii)
similar u p lift pattern offsho re and a re kinematica lly earthquake of November 1975. Rep. 80-8. 46 pp .. Hawaii Inst. of
indisting uishable as far as tsunami excitation is concerned. T h e Geophys . Univ. of Hawaii, Honolulu. 1980.
Chadwick. W. W.. J. R. Smith. J. G. Moore. D. A. Clague. M. 0 . Garcia,
total volume or displaced water is estimated to be - 2.5 km3.
and C. G . Pox, Rathymetry of south flank of Kilauea volcano, Hawaii ,
The slumping area used in the calculation is clos.c to t he size of U.S. Geo /. Survey Invest. Map, MF-2231. 1993.
2800 km2 offshon: bench in the bathymetry [Chadwick et al .. Crosson, R. S., and E. T. Endo, Focal mechanisms and locations of
1993}. earthquakes in the vicinity of the 1975 Kalapana earthquake
From t he comparison of tsunam i excitation and the s ingle- aftershock zone 1970-1979: Implications fo r the tectonics of the south
!lank of Kilauea volcano. Hawaii. Tectonics. I. 495-542. 1982.'
force m odel su ggested earlier from se ismological data we Denlinger, R. P., and P. Okubo, Structure o f the mobile south flank of
conclude that a combinatio n of a faultin g and a large-scale Kilauea volcanoes, Hawaii. J. Geophys. Res.. JOO, 24.499-24,507,
slu mping on the south flank of Ki lauea volcano is the m ost 1995.
appropriate mode l for the 1975 Kalapana earthquake. The Dvorak. J J , An earthquake cycle along the south flank of Kilauea
volcano, Hawaii . J. Geophys. Res.. 99. 9533-954 1, 1994 .
fau lting and slumping are indicative of slope instability o f the
Dvorak, J. LA. T. Okamura, T . T. English . R. Y. Koyanagi. J. S. Nakata,
south flank of K ilauea in res ponse to magmatic or g rav itationa l M . K. Sako, W. T. Tanigawa, and K. M. Yamashita, Mechanical
force s prevailing in the south tlank. This mode l is similar to that response of the south flank of Kilauea volcano. Hawaii, to intrusive
envisaged by Wyss and Kovach (1988]. events along the rift systems. Tectonophysics. 124. 193-209, 1986.
Dvorak. J. J ., F. W. Kle in. and D. A. Swanson. Relaxation of the south
Acknowledgments. We thank the reviewers R. Denlinger and R. Hank after the 7 magnitude Kalapana earthquake, Kilauea volcano,
Crosson for critical reading of the manuscript. valuable comments, and Hawaii, Bull. Seismal. Soc. Am.. 84. 133-141, 1994.
Eissler, H. K and H. Kanamori. A single-force model for the 1975
suggestions that helped us to improve the manuscript. This work was
supported by the National Science .Foundation gra11ts EAR-9303804 and Kalapana, Hawaii, earthquake, J. Geophys. Res .. 92. 4827-4836,
1987.
EAR-9316528 and National Science Council, Taiwan. contract NSC83-
Furumoto. A. S. , and R. L. Kovach. The Kalapana earth quake of
0202-M-008-050T. Contribution 8487 ·of the Division of Geological and
November 29, 1975: An intra-plate earthquake and its relation to
Planetary Sciences, Califo rnia Institute of Technology, Pasadena.
geothermal processes, Phys. Earth Planet. Inter.. 18 . 197-208. 1979.
Cal ifornia.
Geist. E. L., Local tsunami and earthquake source parameters, Adv.
Geophys.. in press. 1997.
References Got, J.-L.. J. frechet. and F. W. Klein, Deep fau lt plane geometry
inferred from multiplet relative relocation beneaJh the south flank o f
Aida, I.. Numerical experiments for the tsunan1i propagation: T he 1964 Kilauea,}. Geophys. Res.. 99. 15.375- 15,386, 1994.
Niigata tsunami and the 1968 Tokachi-Oki tsunami, Bull. Hatori. T., Wave source of the Hawaii tsunamis in 1975 and the tsunami
Earthquake Res. Inst. Univ Tokyo. 47, 673-700. 1969 . behavior in Japan (in Japanese), Zisin. 2. 355-363, 1976.
Aida, I.. Re liabil ity of a tsunami source model perived from fault Hill , D. P.. and J. J. Zucca, Geophys ical constraints on the structure of
parameters, J. Phys. Earth , 26. 57-73 , I 978. Kilauea and Mauna Loa volcanoes and some implications fo r
Ando. M., The Hawaii earthquake ofNovember 29, 1975: Low dip angle se ismomagmatic processes, in Volcanism in Hawaii, vol. 2, edited byMA ET AL MECHANISM OF 1975 KALAPANA EARTHQUAKE 13,167
R. W. Decker, T. L. Wright and P.H. Stauffer. US Geo/. Surv. Pr(Jf Ma, K.-F., K. Satake, and 1-1. Kanamori, The tsunami excited by the 1906
Pap .. 1350, 903-917. 1987. San Francisco earthquake. Bull. Seismol. Soc. Am., 81, 1396-1397.
Hwang, L.-S., D. Divoky, and A, Yuen, Amchitka tsunami study, Rep. l991b.
TC-I 77, 84 pp., Pasadena, Calif, Tetra Tech. Inc.. 1970 Murty. T. S .. Seismic sea waves-tsunamis, Bull. Fish. Res. Board Can,
Hwang, L.-S.. H. L. Butler. and D. Divoky. Tsunami model: 198. 1-337, 1977.
Generation and open-sea characteristics. Bull. Seismal. Soc. Am., 62. Owen, S., P. Segall, J. Freymuller, A. Miklius, R. Denlinger, T.
1579-1 596. l 972a. Arnadottir, M. Sako. and R. Burgmann, Rapid deformation of the
Hwang, L.-S .. H. L. Butler, and D. Divoky, Tsunami generation and south flank of Kilauea volcano, Hawaii, Science, 267. 1328-1332.
propagation. paper presented at 13th International Conference on 1995.
Coastal Engineering, Vancouver, B. C., Can .. July 10-14. 1972b. Satake, K., Inversion of tsunami waveforms for the estimation of a fault
Imamura, F.. N. Shuto, S. Ide, Y. Yoshida, and K. Abe, Estimate of the heterogeneity: Method and numerical experiments, J. Phys. Earth, 3j.
tsunami source of the 1992 Nicaraguan earthquake from tsunami data, 241-254, 1987.
Geophys. Res. Lett., 20, 1515-1518, 1993. Satake, K.. Inversion of tsunami waveforms for the estimation of
Imamura. F , E. Gica, T. Takahashi, and N Shuto, Numerical simulation heterogeneous fault motion oflarge submarine earthquakes: The 1968
of the 1992 Flores tsunami. interpretation of tsunami phenomena in Tokachi-oki and 1983 Japan Sea earthquakes, J. Geophys. Res.. 94,
northeastern Flores island and damage at Babi island. Pure Appl. 5627-5636. 1989.
Geophys , 144, 555-568, 1995. Swanson. D. A.. W. A. Duffield. and R. S. Fiske, Displacement of the
Johnson, J. M, and K. Satake. The 1965 Rat islands earthquake: A south tlank of Kilauea volcano: The result of forceful intrusion of
critical comparison of seismic and tsunallli wave inversions. Bull. magma into the ri1l zones, U.S. Geo/. Surv. Prof Pap., 963, 39. 1976.
Seismal Soc. Am.. 86. 1229-1237, 1996. Tanioka, Y., and K. Satake, Fault parameters of the 1896 Sanriku
Johnson, J. M. and K. Satak:e, Estimation of seismic moment and slip tsunami earthquake estimated from tsunami numerical modeling.
distribution of the April L 1946, Aleutian tsunami earthquake, J Geophys. Res. Lei/. 23. 1549-1552. l 996.
Geophys. Res., J 02, I l,765-1 l, 774. 1997. Tilling, R. I.. R. Y Koyanagi, P. W. Lipman, J. P. Lockwood. J. G.
Johnson. J. M., K.. Satake, S . R. Holdahl, and J. Sauber, The 1964 Moore, and D. A. Swanson. Earthquake and related catastrophic
Prince-William-Sound earthquake: Joint inversion of tsunami and events, island of Hawaii, November 29, 1975: A preliminary report.
geodetic data. J Geophys. Res., /()I. 523-532, 1996. US Geo/. S11n1. Circ., 740, 33, 1976.
Kajiura, K., Tsunami source, energy, and the directivity of wave Wyss, M.. R L. Kovach. Comment on "A single force model for the
radiation. Bull. Earthquake Res. Inst. Univ. Tokyo, 48. 835-869. 1970. 1975 Kalapana. Hawaii, earthquake" by Holly K. Eissler and Hirao
Kawakatsu, H., Centroid single-force inversion of seismic waves Kanamori, J Geophys. Res .• 93. 8078-8082, 1988.
generaled by landslides, J Geophys. Res .. 94. 12.363-12.374, 1989.
Lamb. H .. Hydrodynamics, 738 pp., Cambridge Univ. Press, New York, H. Kanamori, Seismological Laboratory, California Institute of
1932. Technol ogy. Pasadana. CA 91125 .
Lipman, P.W .. J. P. Lockwood, R. T. Okamura. D. A. Swanson. and K. K.-F. Ma. Institute of Geophysics. National Central University,
M. Yamashita, Ground deformation associated with the 1975 Chung-Li 320-54. Taiwan. (fong@sal.gep.ncu.edu.tw)
magnitude-7.2 earthquake and resulting changes in activity of Kilauea K. Satake. Seismotectonics Section, Geological Survey of Japan.
volcano, Hawaii, U S Ceo!. Surv Prof Pap. , 1276,45 , 1985. Tsukuba, lbaraki 305, Japan.
Ma, K.-F., K. Satake and H. Kanamori, The origin of tsunami excited by
the 1989 I.oma Prieta earthquake: Faulting or slumping, Geophys. Res. (Received September 9, 1997; revised Februrary 16, 1999;
lei/ .. 18, 637-640, l991a. accepted March I. 1999.)You can also read
