EERI SPECIAL EARTHQUAKE REPORT -AUGUST 1993scott/GG130/Readings/...Chubu earthquake occurred imme-diately to the south of the 1993,event; the magnitude, fault dimen-)ions, and focal

EERI SPECIAL EARTHQUAKE REPORT -AUGUST 1993

2


of Okushiri Island which lies westof the Hokkaido mainland andsouth-southwest of the epicenter(Figures 3a and 3b). The amountof subsidence initially ranged fromabout 20 cm to 100 cm, but abouthalf of that subsidence has beenrecovered by uplift of the island.Extensive marine seismic surveysby the Japan Maritime SafetyAgency following the earthquakehave revealed extensive imbricatefaulting of the sea floor extendingfrom east of Okushiri Island west-ward to the inferred surface ex-pression of the subduction zone.It is inferred that Okushiri Island islocated on a graben structure with-in the accretionary wedge of theoverriding plate and that this gra-ben structure experienced coseis-mic subsidence. The sea floorstructures discovered during thissurvey resemble those that havebeen found in the sea floor off thePacific coast of Japan. D

~ARTHQUAKE MECHANISM

The earthquake caused subsidence

~ North AmeriCan-*

~ or Okhotsk?) ,cI(Plate f -

HOKKAIDO (

// I

/I

12 July 1993 M:7.8

~

v.JI

Eurasian Plate

The earthquake began on July 12

at 22: 17 local time at latitude 42

47' North, 139° 12' E at a depth

of 27 km. The aftershocks define

a plane about 150 km long (north

south) and 50 km wide (east-west)

dipping eastward at a shallowangle (about 10 ° ) , with most of

the aftershock foci occurring at

depths between 10 km and 20 km

based on preliminary information

from Hokkaido University (Figure

A centroid depth of 1 5 km (a

point source representative of the

overall faulting) was estimated by

Harvard. The seismic moment of5.6 x 1027 dyne cm corresponds tO

an Mw of 7.8. Assuming a fault

length of about 150 km and a fault

width of 50 km based on after

shock data, this moment and fault

geometry indicates an average slipof 2.5 m on the fault plane. The

( ~MT focal mechanism estimatedJ>y Harvard, representing the fault-ing process as a whole, has onenodal plane dipping eastward at24 ° from the horizontal. This

fault plane orientation is compat-ible with the aftershock distribu-tion and consistent with subduc-tion of the floor of the Sea ofJapan beneath northern Japan.However, the source processseems to have been complex, withthe initial rupture occurring on asteeper eastward dipping plane oreven possibly on a shallow west-ward dipping plane (Tanioka et al.,1993; Kikuchi, 1993).

Plate Boundaries 78

()

Pacific Plate

SHI~

\ 2~

The subduction plate boundaryalong the Japan Sea coast ofnorthern Honshu and Hokkaido hasbeen recognized only recently(Figure 2). The 1983 Nihon KaiChubu earthquake occurred imme-diately to the south of the 1993

,event; the magnitude, fault dimen-)ions, and focal mechanism werevery similar to those of the 1993earthquake (Satake, 1985; Kana-mori and Astiz, 1985). The 1940

J~ KYUSHU

48,'--- 68/ // Philippine Sea Plate

0 SOOkm

dFigure 2-Rupture zones of offshore earthquakes in Japan larger than

magnitude 7.4 since 1923. Source: modified from Japan Meteorological.

Agency.

..,-- --,.- , -, ,. -,. ", Shakotan-oki earthquake, which

occurred immediately to the north

of the 1993 event, had a similar

focal mechanism, but smaller mag-

nitude and fault dimensions (Sata-

ke 1986). The occurrence of the

1983 event led to the suggestion

of nascent subduction along the

Japan Sea coast off northern Hon-

I shu (Kobayashi, 1983; Nakamura,,1983) which the 1993 earthquake

tends to confirm. The Sea of

1 ). Japan side of this plate boundary

1" is the Eurasian plate" After the

1983 earthquake, the northern

Japan side of the plate boundary

was initially interpreted as the

North American plate, but it may

instead be the recently proposed

Okhotsk plate (Seno, 1987; De-

Mets et al., 1992)"

EERI SPECIAL EARTHQU:AKE REPORT -AUGUST' 993

Map of Sea Floor Around Okushiri IslandInferred Cross Section of Okushiri Island and

Coseismic Crustal Deformation Model

Okushiri Island

\ i I

\ y»" ~~EPicenter of Hokkaido~~ ~ INanseiOki

~ ~~ ~ 0'""'0'0000000'"0000

1, " .:Y') io!o.'°'","" .--i :; , -~ , -\;~;ijo~*i!

Sea level:~~~~~~~~~~~:::--- O \;:.:-

~

,~a~t:~ apan (

Sea \,

.

~

~'Epicenter

I

I Study AreC"\

--I ESA h,

I ~~r~~t~~;t-"ke;~~i]/6

EurasianPlate North American Plate -'

Main Fault / /

\(i,,~i;,\~;;':

./ Major Active Fault

,,:::::;:::::: Large Scale Landslide Scar":::::::::"

Figure 3 -Sea floor faults inferred from marine surveys after the

earthquake. Source: Yomiuri Shlmbur, Japan MaritIme Safety Agency

RECONNAISSANCE TEAM

A combined EERI-UJNR team sur-

veyed damage and other effects

from the earthquake and tsunami.

The team arrived in Japan on July

18, and conducted field reconnais-

sance and data gathering between

July 20 and 24.

to the field led by Yutaka lida

(PWRI), leader, Kazuhiko

Kawashima (Head of the Earth-

quake Engineering Division, PWRI),Susumu lai (Head, Geotechnical

Earthquake Engineering, P&HRI) S.

Unjoh (PWRI), T. Nakajima (PWRI),

and R. Tsunaki (PWRI).

The acronym UJNR refers to theU.S.-Japan Panel on NaturalResources, a cooperative inter-national program that celebratedits 25th year in 1993. The UJNRPanel on Wind and Seismic Effectsis one aspect of the larger UNJRprogram. Members of the Japan-side Wind and Seismic Panel

arranged outstanding logisticalsupport for the U.S. team, forwhich all members were verygrateful. The UJNR cooperationalso made it possible for the U.S.team members to work closelywith Japanese investigators.

the tsunami report, which was co-authored by E. Bernard, F. Gonza-lez, and D. Sigrist of the U.S. sideof the UJNR team, and HiroichiTsuruya, Port and Harbor ResearchInstitute, Ministry of Transport,and Kenji Kato, University ofTokyo, of the Japan side of theUJNR team, and the landslidesreport, which was authored by E.Harp, USGS, of the UJNR team. D

EERI team members included Les

Youd, Brigham Young University,EERI team leader (geotechnical

aspects); Jane Preuss, Urban

Regional Research (tsunamis, land

use and response planning); Paul

Somerville, Woodward-ClydeConsultants (seismology) ; and

Charles Scawthorn, EOE Inc.

(structural engineering and fire

response). Other members of the

UJNR team included Riley Chung

(NIST), leader; Richard Bukowski

(NIST); Eddie Bernard (NOAA);

Frank Gonzalez (NOAA); Dennis

Sigrist (NOAA); Charles Barnes

(American Plywood Association);David Tyree (American Forest and

Paper Association); Edwin Harp

(USGS), and Peter Yanev, EOE Inc

(NCEER). Several Japanese

engineers accompanied the team

The research, publication, anddistrlDution of this report wasfunded by National ScienceFoundation Grant #BCS-9215158as part of the EERI Learning from

Earthquakes Project.

Complete copies of reconnaissancefireports submitted to EERI are Varchived at the EERCINISEE lIbrary,

ICall EERC at 510-231-9401 forfurther information.

The reports in this special editionof the EERI newsletter wereauthored by the EERI team'members, with the exception of

4

Okushiri Ridge


damage, and few gravestoneswere overturned, suggesting thatthe ground motions in this regionwere not unusually large.

( ~TRONG GROUND MOTION

Strong ground motion recordingswere made at several localities inthe eastern part of southwesternHokkaido and farther to the northnear Sapporo. The largest peakacceleration, slightly less than0.5g, was recorded at Kuromat-sunai which lies midway betweenSuttsu on the Japan Sea coast andOshamanbe on Uchiura Bay. Thatlocality is about 80 km from thenearest point on the fault. Thisrecording had an unusually largeratio of peak horizontal to verticalmotions (a factor of about 5).Suttsu, which is at a similar dis-tance, had peak horizontal accel-erations of about 0.2g and peakdisplacements of about 8 cm. Theduration of strong ground motionwas approximately one to twominutes at most locations within100 km of the rupture surface, asexpected from an earthquake ofthis size. The duration of highfrequency motion was markedlylonger at stations east of theepicenter than at stations to thesoutheast, which may reflect theeffect of rupture directivity fromnorth to south.

Strong ground motions were re-corded on a variety of instrumenttypes by numerous agencies in-cluding the Hokkaido DevelopmentDepartment, the Japan Meteoro-logical Agency I Japan Rail, theMinistry of Construction, and thePort and Harbor Research Institute.There were no strong motion re-cordings from Okushiri Island orthe immediately adjacent south-west coast of Hokkaido. Okushiri

island lies just east of the loweredge of the rupture surface, whichhad a depth of about 20 km. Atthe Aonae cemetery, at the southend of the island, 80% of thetombstones were overturned, andin houses furniture was over-turned, suggesting peak accelera-tions in the range of 0.4g to 0.5g.On the mainland of Hokkaido be-tween Esashi and Motsutanosaki,there was no major structural

1

-Croule

-Youngl-Cohee

0

00,0Q:DoO

00'0O~

( ~~0

:,jII~GI

-GICJ

~

~0

.~0

=

-=GI

~

00 '\

0

§00

00 o~

00

0

0.001 I10 100

Closest Distance (km)

Very preliminary peak horizontalacceleration values without instru-ment corrections are plotted as afunction of closest distance to thefault in Figure 4; these values willchange after the data have beencorrected and checked. Theyarecompared with attenuation rela-tions for subduction earthquakesfor soil sites that are in current usein the United States (Cohee et al.,1991; Crouse, 1991; Youngs etal., 1988). The peak accelerationsrecorded during the Hokkaido-nansei-oki earthquake are generallylarger in the distance range of 70to 1 50 km and attenuate morerapidly than the attenuationrelations. D

Figure 4 -Very preliminary uncorrected peak horizontal acceleration

~ ~' lotted ~gainst c~osest distance t,o the rupture surface, , compare.d with

:1 ttenuat/on relat/ons for subduct/on earthquakes used /n the UnIted

States. Sources: Hokkaido Development Agency, Japan Meteorological

Agency, Japan Rail, Ministry of Construction, Port and Harbor Researchi Institute. ;

5

0

0.

.1

01


TSUNAMIlocations --on the beach, inwrecked automobiles, and in -,homes and other buildings --whichU'are taken to be a rough indicationof tsunami arrival times.

A major tsunami was generatedand, within two to five minutes,extremely large waves engulfedthe Okushiri coastline and thecentral west coast of Hokkaido.As of July 21, 185 fatalities wereconfirmed, with 120 attributed tothe tsunami; this death toll isexpected to rise, as missing per-sons are included among the fatali-ties (Hokkaido Police Headquar-ters) .Property losses have beenestimated at $60 million, dueprincipally to tsunami damage.

killed by salt water. Measure-ments of vertical height above sealevel of the tsunami trace areobtained by a series of measure-ments down to the shoreline, usinga surveyor's staff and level. Theeffects of tides are normallyremoved from such measurements,but during the period of thesurvey, the predicted tidal range inthe study area was only 2-30 cm,referred to mean sea level. Sincethis range is a relatively smallfraction of the observed runup, the

The tsunami was probably gener-ated within the deformation areaas defined by the aftershock pat-tern shown in Figure 1 .Eyewit-ness accounts collected by Y. Tsuji(University of Tokyo) and F. Ima-mura (Tohoku University) indicatethat the tsunami hit the west coastof Okushiri Island almost immedi-ately after the main shock. TheUJNR tsunami survey team found

the electricalclock at an elec-trical power sta-tion had stoppedat 22:23, whichwas 6 min afterthe start of the

earthquake.Along the westcoast, the tsu-nami runupmeasurementswere the high-est from northof Monai. Thevillage of Monaiwas totally des-troyed ( 10 per-sons killed andall 1 2 housesdestroyed) .

Photo: D. Sigrist Runup measure-

ments aroundthe village were 20 m, and in asmall valley north of Monai, therunup was measured at 31 m.Figure 5 shows the survey team inaction in the valley where themaximum runup of 31 m was mea-sured. The staff in the photographis 25 ft high, and the vegetationon both sides of the valley showsevidence of flooding. South ofMonai, tsunami runups between15 and 20 m were measured allalong the coast. Vegetation was Istripped off the hillside, and largeboulders (up to 1 m in diameter)were deposited where the vegeta-tion was flooded. These data are

The Japanese

immediately dis-

patched damageassessment and

survey teams.

Most of these

Japanese teams

were mobilized

and began sur-

veying tsunami

runup by July13. On July 18,

three U.S. scien-

tists joined the

Japanese scien-

tists to comple-

ment the tsuna-

mi survey teams.

A combined US-

Japan team. .worked together F,gure 5 -Runup of 31 m was measured north of Monal.

under the aus-

pices of the U.S.-Japan

Cooperative Program in Natural

Resources (UJNR). (See Recon-

naissance Team, po 4, for names

of team members.)

data presented in this report arenot corrected for the tides. TheUJNR survey team made measure-ments at 28 sites --24 on OkushiriIsland and 4 on the west coast ofHokkaido from July 20-23, 1993.

The primary measurements in thefield surveys were the tsunamivertical and/or horizontal runupvalues, which are the maximumvertical height and/or horizontalextent of flooding. In this report,only vertical runup estimates arepresented. Traces left by the tsu-nami include water marks on build-ings, debris lines along the coast,or vegetation that is damaged or

Another source of information isthe approximate time of tsunamiarrival, as inferred from the time atwhich clocks have stopped due tosaltwater flooding. The UJNRteam conducted an active searchfor such clocks in the tsunamidebris and in the surrounding area'of each of the 28 sites visited.Eight clocks were found in various

6


consistent with the initial wave~rriving from the west of Okushiri,tery near the generation area .

(according to Y. Tsuji) 13 personswere killed and all houses weredestroyed by waves that ran upover 1 O m. The eastern side ofOkushiri Island was less affectedwhere the tsunami runup wasmeasured between 2 and 5 m.Subsidence was observed by Tsujiat Acnae, Okushiri, and Monai,while uplift was observed at InahoPoint. These data again suggestthat Okushiri Island was part of thedeformation area, which is consis-tent with the aftershock data.

(if

peninsula. The UJNR team ob-served that runup values roserapidly again to the 15-20 m levela short distance northeast ofAcnae; this is undoubtedly due tothe absence of breakwaters orsand dunes along this part ofexposed coast.

The extent of damage to Aonae isillustrated in Figure 6. Thebreakwater in the top right corneris 4.5 m. The houses in thecentral part of the photographwere flooded by tsunami wavesthat ran up to a height of 5 m.The area in the top left portion of

~

The tsunami was refracted by theshoaling bathymetry at both endsof the Island. Hardest hit was thetown of Aonae (population 1,600),where the first tsunami waveflooded the southern tip of theisland and the entire first row ofhouses in the harbor area within 4-5 min after the main shock (seeFigure 6) .Tsuji reported (based oneyewitness interviews) that thetsunami arrived from the north-east, with flooding of 3- 7 mthroughout thetown. About 7min after thefirst wave, asecond, largerwave hit fromthe east carryingboats into themain town. Thesecond wave

~ompletelyflooded the firstthree rows ofhouses, and run-up was mea-sured around 5-10 m throughoutthe town. TheUJNR surveyteam found bat-tery-operated Iclocks in this Figure 6- Earthquake, tsunami: and fire damage in Aonae

area that hadstopped at 22:37 and 22:38. At22:40, fires broke out; the combi-nation of a strong northeast windand an ample supply of propaneand kerosene (used for heating)Quickly spread the fire, whichburned throughout the night anddestroyed 340 homes. Autopsiesrevealed that only 2 of the 114deaths in Aonae were caused byfire. This section of Aonae wasthe hardest-hit developed area in

ii -}'Wpite of the fact that a massive,'V4.5 m breakwater and 10 m high

sand dunes were very effective inreducing the runup to 5-10 malong the southeastern tip of the

The tsunami also hit the island ofHokkaido, arriving at Ota Bay with-

in 5 min of themain shock anddestroying five

" homes. TheUJNR team mea-sured runup inthis area at 9 m.Damage was al-so observed bythe UJNR teamat Setana due to6 m runupwaves. Thecoastline fromSuttsu to southof Ota Bay washardest hit withrunup values of5-9 m. Outsidethis area, thetsu-nami

Photo: D. Sigrist intensity tapered

off rapidly, andrunup values fall below 5 m. Theclosest tide gauge to theearthquake was Esashi (Japan.Meteorological Agency), whichrecorded a 2-m wave approximate-ly 10 min after the main shock.Those data suggest that the east-ern portion of the source was be-tween Okushiri Island and Hokkai-do. The tsunami propagated toRussia within 30 min, where 1-4 mtsunami runups were reported byValentin Fedorey (Hydrometl. Af-ter 90 min, the tsunami struck thecoast of South Korea, where B.H.Choi of Sung Kyun Kwan Univ. mea-sured tsunami runup of 1-2 m. O

the photograph had runup valuesof 10 m. The tsunami destroyed aportion of the seawall to the rightof the remaining building in thecenter of Figure 6. Notice that thewave forces transported the bargewith a crane from the harbor tothe center of the photograph .Remnants of homes damaged bythe quake, tsunami, and fire areseen throughout the harbor areatoward the land. The photographswere taken on July 20, 1993.

A similar refraction of the tsunamitools place on the northern point ofOkushiri Island at Inaho where

7


8


stopped. A second and larger

wave arrived after the first waves.inside their houses. Interestingly,

the water velocity was quite low,

although the wave was approxi-

mately 3 meters high (it apparently

passed over the houses) .The first

wave arrived 1 to 2 minutes after

the shaking stopped and did not

recede. The next wave arrived

about 4 minutes later and rushed

over the top of the first.

Warnings and Responses

With the initiation of strong shak-

ing, people reported that theybegan to run immediately for

higher ground (not more than 100m from most residences) .There

was one reported incident, on the

east coast in the vicinity of

Torigashira, of a man who went

back to look between the first and

second waves and never returned .

Five minutes after the earthquake,which occurred at 10: 1 7 pm localtime, the Japan MeteorologicalAgency (JMA) issued a warningthat a major (over 3-meter high)tsunami had been generated. TheJMA transmitted the warning via 7different television and radiostations to local jurisdictions onHokkaido and northern Honshu. Inaddition, N TV issued a notice 1.5minutes after the earthquake thata possible tsunami had been

generated.

Interviews with residents in thenorthern and eastern parts ofOkushiri Island indicate that theyfelt the shaking, looked up, andsaw the first wave arriving. InYamasedomari, the first wavearrived approximately 3 minutesafter the shaking began. However ,a larger second wave arrivedapproximately 1 minute later .

In Aonae, people reported remem-bering the tsunami of 1983, andtherefore, when they felt thestrong earthquake, they knew thatthey needed to run very quickly tothe higher ground (as reported ininterviews with adults, includingone elderly respondent) .The firstwave arrived approximately 5 min-utes after the shaking; the secondarrived about 3 minutes later .

The JMA has six local centers forissuing tsunami warnings: sap-poro, Sendai, Tokyo, Osaka, Fuku-oka and Naha. Each local center isresponsible for issuing a warningsfor its region. Tsunami warnings

pre gra~ed according to three

Ucategorles:

When the shaking occurred, firefighters in Aonae immediatelydrove to the waterfront area towarn residents to evacuate be-cause of a possible tsunami. As

In Yamasedomari, the wave heightwas not high, and the people didnot evacuate, but watched from

.Major tsunami -over 3 m high

.Tsunami warning -1 m to 2 m

.Tsunami alert -0.5 m to 1 m

Table 1: Time-History of Tsunami Warnings

Time Action

22:17 Earthquake occurred.

22:1830

N TV (local television) issued a notice that a possible

tsunami had been generated.

22:22

Table 1 lists the times of tsunamiwarnings issued by Sapporo Dis-trict Meteorological Observatory(DMO) and by the Japan Meteoro-logical Agency (JMA). As indi-cated in the table, five minutesafter the earthquake, SapporoDMO issued a warning that amajor (over 3-m high) tsunami hadbeen generated. The area towhich the major warning wasissued is shown in Figure 7.

Sapporo DMO issued a major tsunami warning for the west

coast of Hokkaido, a tsunami warning for the east coast of

Hokkaido, and a tsunami alert for the north coast of

Hokkaido.

22:24 JMA issued a tsunami alert to mid-coastal sections of

western Honshu.Okushiri Island

Okushiri Island is located near thesource of the earthquake. Eyewit-nesses reported that first small

{ lwaves arrived on the east side of&\lthe island during the earthquake

shaking. The first waves arrived atother parts of the island approxi-mately 2 minutes after the shaking

22:25 Sendai DMO issued a major tsunami warning for the

northern segment of the west coast of Honshu, and a

tsunami alert warning to the next segment to the south-

23:41 JMA issued a tsunami alert for the west coast of Kyushu.

00:12 J~I1A upgraded the warnings for the mid-coastal areas of

Honshu from an alert to a warning .--


deaths were reported in Taiseithey approached the area, the first

incoming wave arrived, and they

retreated up the hill to safety.I

Housing

Hokkaido Mainland

pie seeking refuge on day two, onOkushiri Island, was 1,070. Byday three, the number had risen to1 ,205. Thus, approximately 90percent of displacees requiredshelter. These shelters were setup immediately after the event andat the time of our field investi-gation, 9 days later, approximatelytwo-thirds of those displaced re-mained in the temporary shelters(data from Okushiri Town Hall).

The town of Taisei consists of 11districts, of which 3 coastaldistricts suffered damage. Approx-imately 5 minutes after the earth-Quake, police cars with loudspeak-ers were dispatched to announcetsunami warnings from the streets.Tsunami warning sirens located onpoles adjacent to the fire stationwere not activated because electri-cal power was interrupted at thatlocality by the Quake. Residentsreported being aware that a policecar was issuing announcements,but could not understand themessage because their windowswere closed. Those respondentsreported feeling a strong earth-QUake and preparing to evacuateimmediately by car. One personreported being stuck in traffic andleaving his car to run to safety.

On Okushiri, approximately 300temporary homes in 5 locationsaround the island were being con-structed for the 400 displacedfamilies. Most of the houses wereready for occupancy by July 25.A national law permits occupantsto live in the houses rent free fortwo years; then occupants arerequired to move.

According to preliminary informa-tion obtained from the Laboratoryof Urban Safety Planning in Tokyo,approximately 1 ,495 homes inHiyama Semi-Prefecture were se-verely damaged or destroyed. Theprefecture includes the followingcommunities: Esashi, Kaminakuni,Assahu, Otobe, Kuaishi, Taisei,Okushiri, Setana, Kita Hiyama, andImaganae. In these communities,an estimated 2, 728 people in1,123 families were displaced.

On Okushiri Island, 402 of 680housing units were destroyed. Ofthat total, an estimated 291 weredestroyed by the tsunami and anadditional 108 were initiallydamaged by the tsunami and sub-sequently destroyed by fire. Ofthese homes, 65 were located inan area that was inundated by the1983 tsunami. Assuming an esti-mate of 3.5 persons occupied eachdestroyed home, this destructionwould have displaced 1 ,400 peo-ple. Temporary shelters wereestablished in local meeting housesand schools. The number of peo-

In Taisei, 124 housing units werereported destroyed displacjng apopulation of 291 people. Anadditional number of housing units 4were evacuated because of land-slide damage. On day one, imme-diately after the tsunami, 425people sought shelter. This num-ber dropped to 170 people from62 families the next day. 0

Parts of southwestern Hokkaido

and northern Honshu received a

warning of a major tsunami and

were evacuated. The following

information came from interviews

with residents of Hakodate. The

earthquake knocked out the

power, and a fire truck with a

loudspeaker came by 5 minutes

after the earthquake announcing a

warning to evacuate. No tsunami

impacted this area, and people

were angry at the inconvenience,

especially those who suffered

earthquake damage.

~

Deaths

I

Preliminary casualty estimates, asof July 24, 1993 were 196 fatali-ties, 46 missing, and 34 treatedfor injuries. On Okushiri alone 165were confirmed dead and 39 miss-ing. These totals include 14definite fatalities and 10 stillmissing in the Yo Yo Hotel whichwas buried by a landslide. Ten

Figure 8- The fl.'e was fed by fuel from propane and kerosene tanks, whichwas used for heating and cooking in most homes. Photo: c. Sc8wthorn

10

EERI SPECIAL EARTHQUAK : REPORT -AUGUST 1993

present. Exterior coverings areoften non-combustible stucco orcement board over wood, withcorrugated metal roofing. Largeamounts of exposed wood trim,however, compromise the fire pro-tection. Occupancies are generallycommercial closer to the wharfarea and residential behind (at thebase of the bluff) although manybuildings are mixed occupancies.

carries two 4-m lengths of hardsuction hose equipped with bam-boo strainer baskets. Relative toUS equipment, these fire enginesare smaller in dimensions andcapacity. This smaller size expe-dites passage through narrowerJapanese streets, such as those inAonae. A third fire engine waspresent in Aonae at the time of theearthquake; this engine was inpoor condition, however, and wasparked at the south end of townwhere it was destroyed by thetsunami.

Fire hydrants are located aroundthe town but are not used becausethe watermains are insufficientlysized and pressured to provideadequate water for fire control.Small fires are fought from enginebooster tanks, while the main fireemergency water is stored in un-

The town is protected against fireby a 38-member trained volunteerfire department headed by a full-time professional. The apparatusconsists of two engines of typicalJapanese size and configuration -

each pumper has a 2,000 literbooster tank and carries 10lengths of 20-m long 65-mmdiameter hose. The capacity ofthe pumps is approximately 2,600liters per minute. Each engine also

;t

,(,

Figure 9- Tsunami and fire-damaged areas ';n Aonae, Okush;r; Island. Source: c. ScBwthorn, EQE

11

~~"c-::;t

~

The only known fire ignitionsduring the earthquake occurred inAonae on the southern tip ofOkushiri island. Aonae has apopulation of approximately 500.Most of the town is oriented north-south and sited on or almost onthe beach, only a few metersabove sea level. The rest of thetown is located on a central bluffabout 20 m high where a light-house, the town offices, the firestation, the school, etc., are sited.The lower part of Aonae is denselybuilt-up with narrow streets andtypical building spacings of about3 m. The buildings are generallyone and two story, typically withJapanese wood post and beamconstruction, although some steeland concrete structures were also

EERI SPECIAL EARTHQUAKE REPORT. AUGUST 1993

The initial source of ignition is alsounknown (at this time}; however,villagers told of earthquake shakingturning over all of their furniture, sonumerous ignition sources wereavailable (e.g. cooking and heatingappliances}. At the time of ignition,wind was from the east at about1.5 mlsec with gusts up to about 5mlsec.

documented (Figure 8). Reported-Iy, every time the fire departmen~seemed to be gaining headway, theWfire would flare up again, probablydue to successive involvement ofthese tanks. Additional materialsfueling the spread of the fire wereconsiderable scrap wood in andamong the buildings, and numer-ous vehicles which added gasoline,tires, and flammable interiors tothe conflagration .

derground cisterns sited through

out the town. Individual cistern

capacity is 40,000 liters, which is

accessed through a concrete man-

hole cover .

Shortly after the earthquake. thefire department made a circuit ofthe town looking for fires. Seeingnone and concerned about a possi-ble tsunami, they returned to thefire station. Within a few minutesfollowing the earthquake. thetsunami swept through the lowerarea wrecking many buildings andscattering debris over a wide area.At approximately 10:40 pm thefire department received a citizenalarm of a fire in the lower area. Abrigade of 10 men immediatelyresponded and attempted to reachthe fire by driving down the mainstreet but found the street blockedby debris. They then returned tothe top of the bluff and took asecond route down the southernpart of the bluff .

Firefighting was from hand linessupplied from the pumpers on top ofthe bluff, drafting from the cisterns.Fire progress was south-ward(cross-wind) and relatively slow;suppression efforts signifi-cantlyimpeded fire progress, but thefirefighters were unable to stop thefire. Fire progress was aided byflammables normally stored in eachhome, particularly the fact thatalmost all houses had 490 literelevated kerosene tanks for heatingand propane tanks (20 kg) forcooking. The kerosene tanks werequite likely a principal factor in thefire spread. All such tanks werefound empty after the fire, mosthaving vented safely through thetop vent pipe. The venting wasmost likely caused by radiant heatcausing the kerosene to boil. Eightexploded propane tanks and tworuptured kerosene tanks were

The fire began in a structure abovethe area directly affected by thetsunami, so it likely began as aresult of the earthquake. Theprecise site of initial ignition isunknown, although the approximatelocation is located on Figure 9.

Fire spread was southward atabout 35 meters per hour, withfirefighting on the downwind edge.Two hours into the fire, a secondfire ignited behind the fire line. Atabout 4 am (6 hours after theearthquake), available water fromthe cisterns was exhausted. Citi-zen volunteers assisted in movingthe hose over debris from the blufftop to the port, where the twopumpers drafted from the harbor .At this point, the advancing firefront was about 90 m wide. The ,fire department used equipment to ti)move debris and two buildings,creating a firebreak. Leading fourhand lines from the draftingpumpers, the fire was successfullystopped at about 9 am, savingseveral dozen houses that were inthe path of the advancing fire. D

1~

RETAINING WALLS

Figure 10 -Combination retaining and sea wallnear Oshamanbe that tilted up a few degereesdue to earthquake-generated increase of lateralforces on the wall. Photo: T. i. Youd

""-"'

l'

Tilting of retaining walls occurred Both walls tilted

at several localities. In every port outward by as much

we investigated, quay walls tilted as a few degrees.

from about 10 cm to more than The tilting of the

1 m (see "Liquefaction"). In all of walls generated

these instances liquefaction or fissures in the fill

liquefaction-induced lateral spread that were parallel to

was a major factor in generating the walls. At one

increased lateral forces that locality, in the

caused tilting of the walls. In two Kamiiso District near

instances, however, gravity re- H a k o d a t e , t h e

taining walls tilted, apparently as a cracks were about

consequence of increased lateral 1 m from the wall;

stresses generated in moist, but at the other location

unsaturated backfill. In both of near Oshamanbe,

these instances the walls were the cracks were 2 m

supporting 2- to 3-m high sections to, 3 m from the

of recently compacted highway fill. wall. D


In general, shaking damage tobuildings on Hokkaido and even onOkushiri Island was slight to non-existent. (It should be noted thatdue to the limited field investiga-tion, significant effects may not bereported here.) Tsunami damageon Okushiri Island varied fromcomplete destruction for severelyinundated wood buildings (typical-Iy, all that remained was a con-crete pad), to moderate damagefor concrete or steel buildings wellanchored to their foundations. Forthe latter buildings, tsunami inun-dation resulted in partial or totalloss of contents and cladding onthe first story, stripping thebuilding to its bare structure,leaving the building above theinundation line almost undisturbed.One exception to the general lackof shaking damage was agriculturalstructures. In contrast to most of

())Japan, Hokkaido and OkushiriIsland contain many farm buildingsthat resemble US barns, includingcylindrical silos strikingly similar to,although smaller than, US farmsilos. Japanese silos are typicallybuilt with unreinforced concretehollow-cell masonry unit construc-tion. These types of structuressuffered severely in this earth-quake. About half of the silosobserved on Okushiri Island had

collapsed.

OKUSHIRI-CHO is the main port ofOkushiri Island. The residentialconstruction on the island istypical of Japanese one- and two-story wood frame buildings, withmetal roofing rather than heavierclay tile. Commercial buildings arealso one- and two-story woodframe structures. Shaking damageto Okushiri-cho residential andcommercial construction was mini-

f );\mal, with no building collapses and~only a few chimney collapses.

Chimneys were generally con-structed of lightly reinforcedconcrete.

Figure 11 -Small unreinforced concrete block silo at a farm on Okush,.'i thatsheared and collapsed during the earthquake. Note that the adjacent barn,also constructed of block, is practically undamaged.

Photo: R.P. Orense, Kiso-Jiban Consultants, Tokyo

iJ

~;

UI

turbed. The City Offices are in atwo story stucco-over-wood framebuilding, probably built in the1950's and located approximately1 km from the shore in a narrowriver valley. The building appearsto be founded on firm soils. Therewas no observable damage to thebuilding, other than some loststucco on the west side whichrevealed badly decayed wood. Di-rectl,Y to the west is the two-storyfire station, of similar vintage, but

building was being used to houserefugees, and had sustained veryminor cracking in columns andspandrels. Directly to the west ofthe school is the gymnasium, anapproximately 25 m by 40 m steelframed structure with a barrel archlattice-truss roof. The gym wasbeing used as housing for policeofficers brought in from otherdistricts, and had no observabledamage. Some spalling of stuccocover over concrete was observed

'3

constructed of reinforced concrete.f"\;CSTRUCTURAL DAMAGE: BUILDINGS That buildin.g had no observable\~ damage. Directly to the west and

, "-,-- "- Specific buildings of note in north of the fire station is a stone

Okushiri-cho included the Ferry stairway leading to a hillside one-building, a two story reinforced story wooden Shinto Temple. Atconcrete frame structure with the base of the stairs is.a graniterelatively large, approximately 40 torii, or gateway, whose .stonecm square columns. Located on cross-member had broken. At thethe main pier, the building's first top of the approximately 25 mstory was inundated by the tsuna- flight of stairs is an almost iden-mi, which swept away large glass tical torii of concrete whose cross-windows and all of the interior member is undamaged. The wood-furnishings. The building sustained en temple itself was undamaged.settlement of one column on the Directly across the street to thesouth side, resulting in shear south is the two-story elementarycracking of beams, but was other- school, built in 1970 of reinforcedwise structurally undamaged. The concrete frame construction. Thissecond floor was relatively undis- approximately 30-m by 100-m


on a one-story structure thatconnected the school and gym; thespalling was at the junction of thestructure with the school. To thesouth of the school is the NTTTelephone Exchange building (dis-cussed in "Lifelines"). That struc-ture was also undamaged. Lastly,directly opposite the ferry landinghad been the two-story Yo YoHotel, approximately 25 m by 50m in plan, which was buried undera massive landslide.

the tower to rotate en masse ontothe one-story building. The towercame to rest at approximately a45-degree angle. Our inspectionindicated an apparent bond failureof plain J anchorage bars (approxi-mately 24 mm in diameter) .Alsoon the bluff is the fire station, atwo-story reinforced concretebuilding, that had no apparentdamage. Two schools are locatednorth of the town center, both ofsimilar construction to the Oku-shiri-cho school described above.One of these buildings was twostories high and was undamaged,while the second building wasthree stories high, and sustainedsevere shear cracking in spandrels,

ACNAE was the most severelyimpacted town on Okushiri island.Located at the south tip of Oku-shiri, this fishing community had apopulation of approximately 1.500,

~~~~~

buckled near the base, while theother appeared undamaged. The "\:,amount of cement in the tanks at ;)the time of the earthquake wasunknown. The hopper structureconsisted of a first-story steelmoment frame structure (whichpermitted trucks to enter betweencolumns), with an overhead bracedsteel framework extending to aheight of about 20-m to houseloading equipment and storagefacilities. The framework con-sisted of large cold-formed steelangular column sections, braced byhot-rolled steel angle diagonalbracing. The columns buckled atthe junction of the braced struc-ture with the first-story momentframe, causing the braced struc-ture to tilt about 5 degrees. Asloping 60-m long conveyor struc-ture connected the top of thehopper to processing facilities onthe ground. The conveyor has acylindrical steel pipe-Iike windshield, approximately 1 m in dia- ,meter. The conveyor is vertically f;)supported at mid-span by a steel

IA-frame truss. Due to the large

displacement of the canted hop-

per, the cylindrical steel pipe-Iikewind shield buckled at its junctionwith the A-frame. Repair costs forthe cement plant complex were es-timated at about 20% of replace-ment value.

Figure 12 -Lighthouse at Aonae on Okushiri 'sland that tipped during theearthQuake. The tipping was caused by fracture of the tower structure atthe base along with pull-out of reinforcing steel from the foundation. Ourinspection indicated an apparent bond failure of plain J anchorage bars(approximately 24 mm in diameter).

Photo: R.P. Orense, Kiso-Jiban Consultants, Tokyo

with most residential and commer-

cial buildings located directly

around the port area, only a few

meters above sea level. However ,

local government offices and sqme

housing is located on a bluff above

the port. A lighthouse on the

bluff , a one-story reinforced con-

crete building with a reinforced

concrete tower approximately 8 m

high, failed at the base, allowing

with the cracking extending intobending and shear cracks in thecolumns, and continuing on asshear cracking in the next lowerset of spandrels. The OkushiriCement Plant, in the same generalvicinity, consists of two steel silos,a small office building, and a largehopper for loading trucks with ana~sociated conveyor structure.One of the two steel silos had

)'1

OSHAMANBE is located on Hok-kaido, approximately 80 km fromthe epicentral area. As discussedelsewhere, extensive liquefactionwas observed in this vicinity. Ofparticular interest was the occur-rence of liquefaction at theOshamanbe Nakonosawa Elemen-tary School, a modern complexconsisting of a one-story class-room building of concrete and glu-lam construction, and a steelframed gymnasium. These build-ings were supported on concretecolumns and grade beams and, Idespite extensive liquefaction and

rlateral spreading, sustained only

minor shear cracking in lower wall

panels. O

14


BRIDGES

Few bridges exist on Okushiri andin Hokkaido bridge performancewas generally very good. Nobridges were closed due to struc-tural damage, nor did any bridgessuffer major distress. However ,the approaches to many bridgessettled due to compaction of ap-proach fill, causing vertical separa-tions in the pavement at the junc-tion with the deck, that requiredfilling with asphalt before trafficcould cross the structure, at leastat highway speeds.

1.8 m diameter reinforced concretebridge piers (Figure 13). Thiscracking caused spalling of con-crete in the northern-most pier ,leading to weight restrictions ontraffic crossing the bridge, but thebridge was not closed.

Near Oshamanbe, a bridge tiltedsideways by about 45 cm duringthe earthquake. This bridge wasconstructed in about 1 960 andthen widened later by adding alane to the west side of the bridge,but without widening the founda-tion. Thus, the bridge was subjectto eccentric loading. During theearthquake, the bridge tilted west-ward, in the direction of the eccen-tricity. There was no apparentdifferential settlement of the piers.Liquefaction occurred beneath thebridge, and probably around thepiers reducing lateral resistance tohorizontal loading induced by boththe eccentricity and inertial forcesgenerated by the earthquake. D

Figure 13 -Damaged piers beneatha bridge crossIng the Assabu Riverabout 7 km northeast of Eashi. TheX-pattern of cracks in the piers indi-cate that cyclic flexural stressescaused the damage i';cluding thespalled concrete and buckledreinforcement in the nearest pier.

Photo: T.L. Youd

The most heavily damaged bridge

observed was a six-span, two-lane,

156-m long, 1970 vintage steel

plate girder highway bridge over

the Assabu river, at Assabu-cho,

about 7 km northeast of Esashi.

Lateral spreading was observed in

the south river bank near the

()'t ridge, but did not appear to affectD, ...'y"(he superstructure. Significant

cracking was observed just above

the waterline in the single-column

bers of distribution poles in thetsunami inundation areas, particu-larly on Okushiri, were sweptaway. However, not all poleswere destroyed, and some over-head lines were observed withvegetative debris hanging from thelines, clearly deposited by thetsunami. From this effect, weinfer that poles were probablydestroyed by large debris (e.g,houses) from the tsunami ratherthan directly by the waves. Dam-aged poles and lines were general-ly replaced within ten days of theearthquake. On Hokkaido, severalpoles tilted due to liquefaction ofsupporting soils, causing localoutages until the poles could berighted. Reportedly, local outagesoccIJrred due to transformers beingshaken loose from poles.

WATER AND WASTEWATER -Thewater supply to Aonae is furnishedby an underground pipe, approxi-mately 20 cm diameter, lain alongthe east shore of the island. (Thesource, somewhere to the north,was not determined.) About 1.5 .km north of Aonae the line, awelded steel pipe at that point,crosses a river channel beneath ahighway bridge. The bridge, whichhas about a 6 m clearance abovethe river was undamaged, where-as, the pipeline, with about a 3.5m clearance above the channel,was broken, presumably' by tsu-nami waves. No underground pipebreaks were reported in Aonae.

LIFELINES

~i

HIGHWAYS- In general, highwayperformance was very good, withdamage largely confined to wide-spread settlement at approaches tobridge abutments and settlementof pavements in soft soil areas.Repairs were on-going ten daysafter the earthquake, with a largenumber of damaged pavement sec-tions already repaired.

!\Q4LECTRIC -Electric power doesnot appear to have been signifi-cantly impacted in this event, withthe exception that significant num-

About one mile north of Aonae is a

wastewater treatment plant, set

back from the shore about 300 m

15

Due to our limited field investiga-tions of lifelines, some significanteffects may not have been identi-fied and reported here .


TELECOMMUNICATIONS -TwoRCS's (Remote Control Switcheslexist on Okushiri --one in the townof Okushiri-cho, and other inAonae. The Okushiri-cho RCS islocated in a reinforced concretetwo-story building in the westernpart of the community, approxi-mately 1 km from the shore up ariver valley, on apparently firmsoils. Atop the building is anapproximately 40 m-high free-standing steel truss microwavetower. No damage or problemswere reported for the tower ,microwave, telecom, electrical,HV AC, or other equipment in thebuilding, nor for the building itself .Telecom equipment in the buildingwas well-braced in both directions.Backup diesel generators func-tioned well following the quake.

made to a buried welded steel fuelline with a diameter of about 20 -cm in the southern part of the port ~J'of Okushiri-cho. The line was inthe tsunami run-up zone as well asin an area of apparent extensiveground settlement; either couldhave caused the damage.

Telephone distribution on Okushiriwas severely impacted by the tsu-nami. Lines are carried on thesame poles as the electrical distri-bution, and many were destroyedby the tsunami. Considerable timewas required to splice in new lines.

PORTS AND HARBORS -The portsof Esashi, Aonae, and Okushiri-chowere impacted by the tsunami.Ferry service from Esashi port wasprevented for several days, due toautomobiles and debris in the port. iAonae port had substantial debrisin the port area, as well as generaldestruction of shoreside shops andfacilities. Breakwaters at Esashiand Aonae appeared undamaged.Okushiri-cho port had debris andother problems. In addition, thenorthern concrete breakwater wasdamaged; several breakwater sec-tions were overturned and dis-placed up to tens of meters. Thesections were approximately 3 m -square, 6 m long concrete blocks. ~)Most of the displacement wasapparently caused by tsunamiwave action, although geotech-nical effects may have played arole. O

and consisting of two adjoiningsmall concrete high-bay, one-storybuildings, two approximately 10 mdiameter 5 m high concrete diges-ter tanks, an elevated methane gasholder, and an outfall to the ocean .Moderate settlement of about 30cm occurred around the buildings,which also sustained some moder-ate damage due to pounding be-tween the buildings, and alsobetween the buildings and thedigester tanks. The piping(including significant amounts ofABS piping) and mechanical andelectrical equipment all performedwell, with only minor to negligibledamage. Tsunami inundation inthe area of the buildings was afew tenths of a meter which didn't

cause any significant damage.However, a lower pump room be-neath the larger of the two build-ings was flooded to a depth ofabout 1.6 m apparently due totsunami flooding which entered thebuilding through the outfall.Seawater entered the lower pumproom via the outfall, resulting incontamination of pumps, motors,electrical equipment, etc., suchthat the plant was still shut downten days following the earthquake.

UNDERGROUND FUEL LINES -On

July 23, we noted repairs being

249-260.1983 Alota-Oki earthquake (Mw -7.8) and

its implications for systematics of

subduction earthquakes, Earthquake

Prediction Research 3, 305-317.

Satake, K. (1986). Re-exemination of the1940 Shakotan-oki earthquake and the faultparameters of the earthquakes along the

eastern margin of the Japan Sea, Phys.Earth. Planet. Interiors 43, 137.147.

Cohee, B.P., P.G. Somerville and N.A.

Abrahamson (1991 ). Simulated groundmotions for hypothesized Mw = 8

subductio.n earthquakes in Washington and

Oregon, Bull. Seism. Soc. Am. 81, 28-56;printer's typos corrected p. 2529. Seno, T ., C.R. DeMets, S. Stein, and D.F.

Woods {1987). The movements of the

Okhotsk plate {abstract in Japanese),Seism. Soc. Japan Program and Abstracts

1987 (1), p. 87.

Kikuchi, M., (1992). The magnitude 7.8Hokkaido Nansei Oki earthquake of July 1 2,

provisional analysis of broadbandteleseismic seismograms, Yokohama CityUniversity, Seismology Report No.23

(Appendix).Crouse, C.B. (1991). Ground motionattenuation equations for earthquakes on

the Cascadia subduction lOne, Earthquake

Spectra 7.201-236.

Kobayashi, y (1983). On the initiation of

subduction plates. The earth monthly.

5,510-514 (in Japanese). Tanioka, Y., L. Ruff and K. Satake, (1993),The puzzling rupture process of the July 12,

1993 southwest Hokkaido earthquake,

unpublished manuscript.Nakamura, K. (1983). Possible nascenttrench along the eastern Japan Sea as the

convergent boundary between the Eurasian

and North American plates, Bull. Earthquake

Research Institute, 58, 711-722.

DeMets, C. A test of present-dey plategeometrics for northeast Asia and Japan, J.

Geophysical Research, in press.

Youngs, R.R., S.M. Day and J.L. Stevens

(1988). Near fie~d ground motions on rock .jffor large subduction earthquakes, ASCE ~Geotechnical Publication No.20, 445-462(rock); and Washington Public Power Sup-ply System (1988). Response to NBCQuestions 230.1 and 230.2 (soil).

Hata, Mitsuo; Segawa, Hidoyoshi; andYajima, Junkieki, (1982), Geology of South-

West Hokkaido; Okushiri Island; GeologicalMap (series and affiliation in Japanese),

Scale 1 ;50,000.

Satake, K. (1985). The mechanism of the

1983 Japan Sea earthquake as inferred

from long-period surface waves and

tsunamis, Phys. Earth. Planet. Interiors 37 ,Kanamori, H. and L. Astiz (1985). The


Where liquefiable deposits hadaccumulated in the southern partof Hokkaido, surface effects ofliquefaction were generally evi-dent. These effects included sandboils, fissures, lateral spreads,ground oscillation, loss of bearingstrength, ground settlement, andbuoyant rise of manholes andtanks. The areas affected includedseveral alluvial river valleys on thewest coast of southern Hokkaido,alluvial plains on the east coastaround Uchiura Bay, and infills atseveral ports. These effectscaused varying amounts of dam-age to constructed works, but inseveral notable instances, well-built structures survived theoccurrence of liquefaction with no

significant damage.

Figure 15 -Anchored bulkheads tilted seaward at all of the ports we visitedon the southern peninsula of Hokkaido. This bulkhead at the Port of Hako-date was one of the more severely damaged, with horizontal displacementsat the top of the wall in excess of 1 m. A major fissure of similar widthparalleled the wall about 10 m inland (along the line of anchor points?).Widespread ground settlements and eruptions of sand boils indicate thatcompaction and liquefaction of fill materials were a principal cause ofdamage. Photo: T.L. Youd

! At each of the ports we visited,

~uay walls had tilted seaward from

lf10 cm to 20 cm at ground surface

with one notable instance of about

1 m of displacement. Ground fis-

sures, settlements and sand boils

disrupted pavements and storage

areas behind these walls (Figure

15).

Where highway grades crossedliquefiable areas, the road surfacecommonly differentially subsidedgenerating wavy pavement withwave-lengths usually severalmeters in extent (Figure 16). Thistype of disturbance was noted inseveral areas on both sides ofsouthwestern Hokkaido; at eacharea we specifically examined,sand boils, ground fissures andother evidences of liquefaction haddeveloped near the roadway.

rl

Several bridges crossed rivers andflood plains where liquefactioneffects were evident. In a fewinstances, abutments had rotatedinward as a consequence of in-creased lateral pressures due toliquefaction and lateral spread. Inthese instances abutments jammedthe bridge girders causing minordamage to the seating and/or tothe bridge girders. In severalinstances, however, the bridgeswere undamaged indioating thatthe foundations and superstruc-

,Figure 16 ~ View of Highway 5 south of Oshamanbe where vertical waves,(pJWith troughs as deep as 0.6 m, disrupted traffic. Similar waves developed~ at several other localities. Nearby fissures and sand boils at each site

examined indicate that liquefaction contfl"buted to the highway disruption.Photo: R. P.'Orense, Kiso-Jiban Consultants, Tokyo

EERI SPECIAL EARTHQUAKE REPORT -AUGUST' 993

oil storage tank at an industrialfacility popped mostly out of th-ground (Figure 18). That tank was"7 .2 m long and 1.8 m diameter.The tank had been about 1/3 full ofoil at the time of the earthquake.

Flood control dikes along rivers onthe west side of southwestern Hok-kaido commonly cracked and settledseverely in areas affected by lique-faction. This type of dike damagehas been a common occurrenceduring several recent Japaneseearthquakes. The dikes were up toseveral meters high and paralleledboth sides of the rivers in lowlandvalleys that cut through the gener-ally mountainous terrain. Cultivatedfields lay beyond the dikes in theprotected areas. The dike settle-ments commonly involved longitu-dinal splitting of the embankment atthe crest and cracking and bulgingof the ground surface along theflanks of the dikes. The mode offailure appears to have been loss ofbearing strength in the liquefied SOiabeneath the embankments withpenetration of the dike into thesoftened soil and lateral and upwardthrusting of the foundation soils oneither side of the dike. Sedimentcompaction may also have added tothe subsidence of the dikes. D

~ foundation support.I Oil storage tanks at

this and other portswere undamagedeven though groundsettlements andsand boils occurredin the near vicinityof the tanks. Onetank we specificallyexamined at the Fer-ry Port at Hakodatehad oscillated hori-zontally relative to

Figure 17 -The ground fissures and sand boils the ground with anshown in the foreground continue beneath the bridge amplitude of aboutwhere even larger fissures and sand boils developed. 10 cm to 20 cm .Free-field lateral spread displacemen~s in ~he n~ar The ground aroundvicinity were as large as 1 m; the brIdge, Includ,ng the tank had settledthe piers and abutments, however, were undamaged. and sand boil depo-The bridge crosses the Toshibetsu River about 5 km sits covered thesouth of Setana. ground surface.

Photo: R.P. Orense, Kiso-Jiban Consultants, Tokyo The tank, which

tures had sufficient strength to was apparently founded on piles,resist ground displacements and was level and undamaged .other deleterious effects ofliquefaction (Figure 17). ' --.,-"---

Locally, several water and sewerlines were fractured or severed dueto differential ground displacementsgenerated by liquefaction. How-ever, most liquefaction occurred innon-populated areas where under-ground pipelines had not been con-structed. The most spectaculareffects of liquefaction to appur-tenant pipeline facili-ties was the rise of ,'Y- --manholes relative to ,

the ground surface,either as a conse-quence of buoyantrise of the structureor settlement of theground. Theseeffects were mostprominent in the cityof Oshamanbe.

Many houses and other small- tomedium-sized buildings were locatedin areas of surface disturbances dueto liquefaction. The foundationwalls or slabs of several buildingswere pulled apart in extension orvertically offset. In nearly allinstances where this type of dam-age occurred, the foundations wereunreinforced or poorly reinforced.Conversely, several buildings onwell-constructed and well-reinforcedfoundations survived the groundshaking and occurrence of liquefac-tion with no significant damage.Several of these structures werefounded on piles with the groundaround the structures subsiding andfracturing, but the structureremaining in place and undamaged.In other instances, buildings onshallow, but well-reinforced andwell-tied-together footings alsoremained undamaged.

Several buried gaso-line or oil storagetanks also rosebuoyantly by a fewcentimeters at gaso-line stations, crack-

ing overlying pave-m.ents and pipelineconnections. One

Figure 18 -Ot1 storage tank that floated out of theground at an industrial fact1ity about 6 km north ofEsashi. Liquefaction of backft11 around the tan/4;:aapparently allowed the tank and cover slab toW'buoyantly rise and tip. The tank was about 1/3 fullat the time of the earthquake.

A reinforced-concrete silo for

storage of cement at the Port of

Hakodate settled and tilted about

2.5° as a consequence of loss of Photo: T.L. Youd

18


to scramble to safe-

ty after being briefly

trapped within the

remains of the hotel.Landslides were triggered through-out Okushiri Island as well asthroughout much of the southwest-ern Hokkaido. The slides consistedmainly of rock falls and rock slideson steep ( > 50. ) natural slopes andengineered cut slopes. Volumesranged from several cubic meters tothe 800,000 mJ rock slide thatburied the Yo Yo Hotel at OkushiriPort (Figure 19). Numerous slumpsoccurred in engineered fill alonghighways and railways.

The hotel was lo-cated just seawardof a main road thatwas aligned alongthe base of theslope. The pre-earthquake slopeinclination wasabout 60 0, parallel Figure 19 -Rock slide near the ferry terminal into the slide surface Okushiri City that buried the Yo Yo Hotel.that was subse-quently exposed by the failure. Theheight of the rock slide from toe tocrown is approximately 100 m, its

along the coast of the island whereextremely steep slopes of volcanicbreccia, basalt, and tuffaceous

sandstone shed debris thatwas shaken apart along highlyweathered fractures. Therock debris blocked coastalhighways in many places. Atmany of these localities, thefalling rock destroyedsegments of concrete barriersthat had been constructed forrockfall protection. At severallocalities, wire nets that hadbeen draped from the cliffshad been torn and penetratedby large rock fragments.These nets had been placedto control the fall of rockdebris.

Rock falls and slides onOkushiri Island

The largest landslide triggeredby the earthquake occurredabout 70 m from the ferryterminal in Okushiri City. This800,000 mJ slide consisted ofPliocene tuff, tuffaceous11andstone, and conglomerate

~ Hata et al., 1982). The slideburied the yo Yo Hotel andseveral vehicles. Two largekerosene storage tanks at thenorthern edge of the slidewere also damaged. The fatal-ities are estimated at 35-40 Figure 20 -Rock fall along seacoast north ofpeople, but these numbers are Setana leaves pinnacle virtually cleaved in half.uncertain because the hotelrecords, as well as the hotel staff,were trapped in the debris. A fewguests on the upper floors were able

width is about 800 m, and itsthickness ranges from 20 m to 40m. The slide's morphology suggests

that its motion waslargely translationaland parallel to its60° scarp. Theslide mass movedabout 40 m with the

11 topmost part staying! largely intact with

the distal portiondilating and spread-ing as the materialunderwent crushingfrom the weight of

! the overlying debris.

Landslides on SouthwesternHokkaido

~

The occurrence of rock falls androck slides triggered by the earth-quake on southern Hokkaido Penin-sula was similar to that on OkushiriIsland. These types of failuresappeared to be most numerous onthe western coast of the peninsulawhere the steepest slopes exist. Ason Okushiri Island, most of the rockfalls were generated from andesiticvolcanic breccias which stand ascoastal cliffs and erosional pinnacleswith slopes of 50 o to vertical.

Several of these pinnacles along theocean shore north of Setana were

~~

Figure 21- Talus shed collapsed by rock fall alonghighway 229 north of Setana. Numerous rock falls

All photos this page by E. Harp, USGS and slides developed

19


virtually cleaved in half by rockslides along widely space fracturesurfaces within the rock (Figure20).

r

About 1 5 km north of Setana, on

highway 229, one of the sections

of a talus shed collapsed from the

impact of a large rock fall (Figure

21 ). The shed was constructed as

a cast-in-place concrete tube. The

rock fall contained large pieces of

andesitic breccia, some greater

than 4 m diameter. The total

volume was several hundred cubic

meters and originated from a near

vertical cliff about 60 m above the

highway.

Figure 22 -Slump north of Oshamanbe on highway 5 within engineered fill

in highway and ral1road embankments resulted in overturning of two trucks

and one car; only one truck remained when this photo was taken.Photo; R.P. Orense, Kiso-Jiban Consultants, Tokyo

About 10 km northwest of Osha-manbe, on highway 5, a principalarterial highway, a slump de-veloped in the engineered highwayfill which overlies marsh deposits.An adjacent railroad fill was alsodisturbed by slumping at thislocality. This failure of several

thousand cubic meters of soil mayhave been induced by liquefactionor shaking-induced softening of the

underlying marsh deposits. The

failure caused the overturning of ~

two trucks and a car (Figure 22). LJ ~

EARTHQUAKE ENGINEERING

RESEARCH INSTITUTE

499 14th Street, Suite 320

Oakland, CA 94612-1902

RETURN POSTAGE GUARANTEED

FIRST CLASS

Postage Paid at

Berkeley I CA

Permit No.204

EARTHQUAKE ENGINEERING

RESEARCH INSTITUTE

PRESIDENT

Lloyd S. Cluff

PAST PRESIDENT

J. Carl Stepp

VICE-PRESIDENT

William T. Holmes

SECRETARY-TREASURER

Chris D. Poland

BOARD OF DIRECTORSThomas L. Anderson

Christopher ArnoldA. Gerald Brady

Lloyd S. CluffC. B. Crouse

Phillip L. GouldWilliam T. Holmes

Chris D. PolandJ. Carl Stepp

(

EXECUTIVE DIRECTOR

Susan K. Tubbesing

20

Documents

EERI SPECIAL EARTHQUAKE REPORT -AUGUST 1993scott/GG130/Readings/...Chubu earthquake occurred imme-diately to the south of the 1993,event; the magnitude, fault dimen-)ions, and focal