J Seismol (2011) 15:19–41DOI 10.1007/s10950-010-9203-x

ORIGINAL ARTICLE

A strong-motion database from the Peru–Chilesubduction zone

Maria C. Arango · Fleur O. Strasser ·Julian J. Bommer · Ruben Boroschek ·Diana Comte · Hernando Tavera

Received: 6 March 2010 / Accepted: 29 July 2010 / Published online: 21 August 2010© Springer Science+Business Media B.V. 2010

Abstract Earthquake hazard along the Peru–Chile subduction zone is amongst the highest inthe world. The development of a databaseof subduction-zone strong-motion recordings is,therefore, of great importance for ground-motionprediction in this region. Accelerograms recordedby the different networks operators in Peru andChile have been compiled and processed in a uni-form manner, and information on the source pa-rameters of the causative earthquakes, fault-plane

M. C. Arango · J. J. Bommer (B)Department of Civil and Environmental Engineering,Imperial College London, London SW7 2AZ, UKe-mail: j.bommer@imperial.ac.uk

F. O. StrasserSeismology Unit, Council for Geoscience,Private Bag X112, Pretoria 0001, South Africa

R. BoroschekDepartment of Civil Engineering,University of Chile, Blanco Encalada 2002,Santiago, Chile

D. ComteDepartment of Geophysics, University of Chile,Blanco Encalada 2002, Santiago, Chile

H. TaveraSeismology Department,Geophysical Institute of Peru, Calle Badajoz 169,Urb Mayorazgo IV Etapa, Ate, Lima, Peru

geometries and local site conditions at the record-ing stations has been collected and reviewed toobtain high-quality metadata. The compiled data-base consists of 98 triaxial ground-motion record-ings from 15 subduction-type events with momentmagnitudes ranging from 6.3 to 8.4, recorded at59 different sites in Peru and Chile, between 1966and 2007. While the database presented in thisstudy is not sufficient for the derivation of anew predictive equation for ground motions fromsubduction events in the Peru–Chile region, it sig-nificantly expands the global database of strong-motion data and associated metadata that canbe used in the derivation of predictive equationsfor subduction environments. Additionally, thecompiled database will allow the assessment ofexisting predictive models for subduction-typeevents in terms of their suitability for the Peru–Chile region, which directly influences seismichazard assessment in this region.

Keywords Peru–Chile subduction zone ·Strong-motion database · Ground-motionprocessing · Site classes · Source and pathparameters

1 Introduction

The development of a strong-motion databasefrom subduction events along the Peru–Chile

20 J Seismol (2011) 15:19–41

trench is an essential step for ground-motion pre-diction in this region as well as other subductionzones in the world where there is a significanthazard from earthquakes along the interface be-tween the subducting and overriding plates andwithin the subducting slab. The tectonic setting ofthe Peru–Chile region is characterised by the sub-duction of the relatively young (∼50 Mya) Nazcaplate beneath the continental South Americanplate (Fig. 1), which takes place at a conver-gence rate of 7–9 cm/year in the N78◦ E direc-tion (DeMets et al. 1990). Along the Peru–Chilesubduction zone, two main types of seismicitycan be identified: firstly, earthquakes occurringat the seismically coupled interface between theNazca and South American plates (interfaceearthquakes) and secondly, seismicity related tothe zone of extension in the interior of thedescending Nazca plate (intraslab earthquakes).The Peruvian–Chilean subduction zone has fre-quently ruptured in great (M ≥ 8.0) destructiveearthquakes during the last centuries, many ofwhich have been thrust-faulting events occurringalong the interface between the Nazca and SouthAmerican Plates, although there have also beena number of damaging intraslab events. An ex-ample of the latter is the great event (MW 8.1)that occurred on 25 January 1939, near the cityof Chillán, which killed approximately 28,000 peo-ple and is amongst the most damaging eventsthat have occurred in the seismic history of Chile(Beck et al. 1998).

The occurrence of major subduction eventsalong the Peruvian–Chilean subduction zone isquite frequent, with 17 events of magnitude MW ≥7.5 registered during the last 50 years. In centralChile, the historical record of earthquakes startswith an event of magnitude M 9.4 in 1575 followedby great events in 1647 (M 8.4), 1730 (M 8.2),1822 (M 8.4) and 1906 (M 8.3; Comte et al.1986). For instance, the great Valparaiso event of17 August 1906 (M 8.3) caused widespread dam-age in central Chile and claimed thousands oflives. In recent years, the central Chile seg-ment ruptured in a large interface earthquake on3 March 1985 (MW 8.0), which killed 177 peo-ple and caused extensive damage in the cities ofValparaiso and Viña del Mar. This latter eventruptured along a previously identified seismic gap

�Fig. 1 Tectonic setting and distribution of seismicity alongthe Peru–Chile subduction zone. Seismicity corresponds tothat reported in the EHB Bulletin (Engdahl et al. 1998) forthe period 1960–2006 (International Seismological Centre2009). The width and direction of the cross-sections ofseismicity are indicated by the rectangles on the map

in central Chile with high probabilities of re-currence for a large earthquake (e.g., Nishenko1985). The north-central Chile segment of sub-duction has also ruptured in great events in 1922(MW 8.7), 1943 (MW 8.2) and 1995 (MW 8.0).The south-central Chile segment, between 35◦ and37◦ S, is another identified seismic gap, referredto as the Concepción–Constitución seismic gap,which has been extensively studied following the1939 Chillán event. This segment ruptured in agreat (M 8.5) earthquake in February 1835, whichcompletely destroyed the city of Concepción. Thelast great earthquake in this region was the mag-nitude MW 8.8 interface event that occurred on27 February 2010, whose epicentre was located at100 km from the city of Concepción. Preliminaryreports indicated that this latter event had a deathtoll of more than 500 people and caused extensivedamage to the cities of Concepción, Arauco andCoronel, affecting over two million people. Southof this region, between 37◦ and 46◦ S, a greatMW 9.5 underthrusting event occurred on 22 May1960, causing 1,660 deaths and leaving two millionpeople homeless. This event, the largest instru-mentally recorded in the world during the twen-tieth century, had an estimated rupture length ofabout 1,000 km (Cifuentes 1989) and also induceda tsunami that spread across the Pacific.

In the southern Peru and northern Chile seg-ment of subduction, major interface events oc-curred in 1868 (southern Peru) and 1877 (northernChile) with magnitudes estimated between 8.5 and9.0 (Lomnitz 2004; Kausel 1986; Dorbath et al.1990). This region had been identified as a seis-mic gap with a high potential of occurrence of agreat earthquake (Lomnitz 2004; Kelleher 1972;Nishenko 1985; Comte and Pardo 1991; Delouiset al. 1996). The southern part of this regionruptured in a large interface event in July 1995(MW 8.0), which occurred south of the rupturezone of the 1877 event. Large intraslab-typeevents have also occurred in northern Chile in

J Seismol (2011) 15:19–41 21

Section 1-1’

-80o -76o -72o -68o -64o -60o

-35o

-30o

-25o

-20o

-15o

-10o

Lon E

Lat N

Seismicity 1960 - 2006 (EHB Catalogue)

Depth of events0-75 km 450-525 km75-150 km 525-600 km150-225 km 600-675 km225-300 km 675-750 km300-375 km >750 km375-450 km

1

Section 2-2’

Section 3-3’

0 200 400 600 800 1000

0

100

200

300

Distance from the trench [km]

Hyp

ocen

tral

Dep

th [k

m]

0 200 400 600 800 1000

0

100

200

300

Distance from the trench [km]

Hyp

ocen

tral

Dep

th [k

m]

0 200 400 600 800 1000

0

100

200

300

Distance from the trench [km]

Hyp

ocen

tral

Dep

th [k

m]

0 200 400 600 800 1000

0

100

200

300

Distance from the trench [km]

Hyp

ocen

tral

Dep

th [k

m]

0 200 400 600 800 1000

0

100

200

300

Distance from the trench [km]

Hyp

ocen

tral

Dep

th [k

m]

0 200 400 600 800 1000

0

100

200

300

Distance from the trench [km]

Hyp

ocen

tral

Dep

th [k

m]

Section 5-5’

Section 4-4’

Section 6-6’

1’

2

2’

3’

1

3

4’4

5’5

6’6

NAZCA PLATE

PACIFICOCEAN

78 mm/year

Peru

SOUTH AMERICANPLATE

80 mm/year

Chile

Peru-Chile Trench

Events used in this study

Relative plate motions

22 J Seismol (2011) 15:19–41

December 1950 (MW 8.0) and 13 June 2005 (MW

7.8). The northern part of this seismic gap rup-tured in a great interface event on 23 June 2001(MW 8.4), along the rupture area associated withgreat 1868 southern Peru event. The 2001 eventhad a death toll of 80 causalities and caused se-vere damage in the cities of Ocoña, Arequipa,Tacna and Moquegua. Along the central region ofPeru, the subduction processes have caused greatearthquakes in 1746 (M 8.5), 1940 (MW 8.1), 1942(MW 8.0), 1966 (MW 8.1), 1970 (MW 7.8), 1974(MW 8.1), 1996 (MW 7.7) and 2007 (MW 8.0), caus-ing thousands of deaths. The Central Peru seg-ment of the subduction zone, between the ruptureareas of the 1974 (MW 8.1) Lima event and the1996 (MW 7.7) Nazca event (Tavera and Bernal2005), had also been identified as another seismicgap. This gap last ruptured in a MW 8.0 event on17 August 2007 in the Pisco region of CentralPeru, causing 595 deaths and extensive damage inthe cities of Pisco, Chincha and Cañete (Taveraet al. 2008).

In view of the threat that both interface andintraslab-type events pose to the Peru–Chile re-gion, the compilation of a strong-motion databaseof subduction events that can be used as the basisof future ground-motion prediction studies is ofprime relevance. In Chile, the first strong-motioninstruments were deployed in the 1970s by theCivil Engineering Department of the Universityof Chile (RENADIC network), which recordedthe 3 March 1985 (MW 8.0) Valparaiso earthquakeand associated aftershocks amongst other events.Presently, the RENADIC network consists of20 analogue and 15 digital stations installed inNorthern and Central Chile. A second network(DGF-DIC) was deployed by the Departmentsof Geophysics and Civil Engineering of theUniversity of Chile and the Swiss SeismologicalService as a part of a project to study the north-ern Chile seismic gap and has been in operationsince 2001. The DGF-DIC network consists of11 digital instruments installed in Northern Chile,from Arica to Antofagasta. These two networkshave recorded several large events, amongstothers those on 13 June 2005 (MW 7.8) and 23June 2001 (MW 8.4). The Geophysical Institute ofPeru (IGP) deployed the first strong-motion in-struments in Lima, which recorded the 1966, 1970,

1971 and 1974 Peruvian events. Currently, strong-motion networks in Peru are operated by IGP, theJapan–Peru Centre for Seismic Research and Dis-aster Mitigation (CISMID), the South AmericanRegional Seismological Centre CERESIS, theCatholic University of Peru (PUCP), and thePeruvian state water company (SEDAPAL).Recent significant events recorded by these net-works include the 23 June 2001 (MW 8.4) and 15August 2007 (MW 8.0) Pisco event.

This paper presents the work performed inorder to develop a database of strong-motionrecords from events along the Peru–Chile subduc-tion zone and associated information (metadata).The strong-motion data recorded by the differentnetworks operators in Peru and Chile have beencompiled and processed in a uniform manner,and information on the source parameters ofthe causative earthquakes, fault-plane geometriesand local site conditions at the recording stationshas been collected and reviewed. Earthquake pa-rameters from different reporting agencies andpublished studies were examined to define re-liable source parameters, fault-plane geometriesand distance metrics. Additionally, geological andgeotechnical information at the recording siteswere collected from different sources, and siteswere classified according to various schemes.

2 Database description

The compiled database consists of 98 triaxialground-motion recordings from 15 subduction-type events with moment magnitudes rangingfrom 6.3 to 8.4, recorded at 55 different sites inPeru and Chile, between 1966 and 2007. Theseaccelerograms have been made available by localnetworks in Chile and Peru including the NationalAccelerographic Network of Chile (RENADIC,23 records), the DGF-DIC network jointly oper-ated by the Departments of Geophysics and CivilEngineering of the University of Chile (sevenrecords), the IGP (eight records) and the CISMID(12 records). Additionally, strong-motion recordsfrom the 1985 Valparaiso (Chile) sequence andfrom the 1966, 1970, 1971 and 1974 Peruvianevents available at the COSMOS Virtual DataCentre (http://db.cosmos-eq.org) were also in-

J Seismol (2011) 15:19–41 23

-84o -80o -76o -72o -68o

-36o

-32o

-28o

-24o

-20o

-16o

-12o

-8o

Lon E

LatN

INTERFACEINTRASLAB

PERU

CHILE

13-06-2005 [MW 7.8]

23-06-2001 [MW 8.4]

31-05-1970 [MW 8.0]

17-10-1966 [MW 8.1]

03-03-1985-A [MS 6.4]

15-10-1997 [MW 7.1]

03-03-1985 [MW 8.0]

09-04-1985 [MW 7.1]

04-03-1985 [MW 6.3]

25-03-1985 [MW 6.4]

05-01-1974 [MS 6.6]03-10-1974 [MW 8.1]

09-11-1974 [MW 7.1]

15-08-2007 [MW 8.0]

07-11-1981 [MW 6.9]

-78o -76o -74o -72o -70o

-18o

-16o

-14o

-12o

-10o

SEDAPALIGPCISMIDCERESISPUCP

ANC

NNA

MAY

CDL-CIPCAL

PUCPRINCER

ANR

MOL

CSM

ANC

PCNICA2

TAC1TAC2

MOQ1MOQ2

MOQ3

AQPAQP-2

ZAR

LMO

GEO

a

-70o -68o

-22o

-20o

-18o

RENADIC

DGF-DIC

MEJI

CALA

LOA

TCP

PICA

IQU

PIS

CUY

PUTRE

ACOPOCO-1

IQUIIQC

ACA

ARIEPOCO-2

b

-74o -72o -70o

-36o

-34o

-32o

RENADIC ILLA

VIL

SFDO

LIGPAP

ZAP

LLAY

LLO

SANT

MEL

CHI

CAU

TALCCON

ILO

HUA

COL

PICH

RAP

QUIN

SFELISIVMAR

VENVALM UTFSM

END

c

a

b

c

Fig. 2 Location of the strong-motion stations used in thisstudy. The panels show the stations located in a Central andSouthern Peru; b Northern Chile and c Central Chile. The

locations and focal mechanisms of the events contributingdata to this study are also shown

24 J Seismol (2011) 15:19–41

cluded in this database (48 records). The locationof the strong-motion stations operating along thePeru–Chile subduction zone, from which record-ings are presented in this study, is shown in Fig. 2.The majority of the data from these agencies havebeen released in unprocessed format; however, ina few cases, strong-motion records to which somelevel of processing has already been applied werealso included. All strong-motion data included inthis database are from either free-field stations orinstruments at the base of structures, at a totalof 59 sites. In the context of this study, free-fieldrecordings are defined as those obtained at sta-tions in small shelters, isolated from any buildinginfluence. The other recordings are obtained frominstruments at the basement of structures up to

three storeys in height, although five recordingsobtained at stations located at the basement ofstructures with more than three storeys were in-cluded in this database.

Figure 3 displays the distribution of the data interms of magnitude, distance, focal depth, eventtype and National Earthquake Hazards Reduc-tion Program (NEHRP) site class. The methodol-ogy used in the determination of the seismologicalparameters, computation of distance metrics andassignment of sites is discussed in the followingsections of this paper. Overall, all strong-motiondata available are from moderate-to-large events(6.3 ≤ MW ≤ 8.4) recorded at distances of about25–420 km from the fault plane. Approximatelyhalf of the entire dataset was recorded at short dis-

Fig. 3 Distributionof the dataset in termsof magnitude, distance,focal depth, event typeand NEHRP site class

20 30 100 200 3005

6

7

8

9

Rupture distance, Rrup [km]

Mom

entm

agni

tude

,MW

INTERFACE INTRASLAB

20 30 100 200 300

0

20

40

60

80

100

120

Rupture distance, Rrup [km]

Foc

alde

pth

[km

]

20 30 100 200 3005

6

7

8

9

Rupture distance, Rrup [km]

Mom

ent m

agni

tude

,MW

NEHRP A&B NEHRP C&C/D NEHRP D&D/E

20 30 100 200 3005

6

7

8

9

Rupture distance, Rrup [km]

Mom

entm

agni

tude

,MW

INTERFACE INTRASLAB

J Seismol (2011) 15:19–41 25

tances (Rrup ≤ 100 km), and consequently, a sig-nificant number of the ground motions are of largeamplitudes; the level of peak ground acceleration(PGA) recorded during these events varies withina range of approximately 20–700 cm/s2. Similarly,most of the data included in the dataset comefrom events with magnitudes MW 8 ± 0.3 andMW 7 ± 0.2. The distribution of focal depth withrespect to rupture distance for recordings fromboth interface and intraslab-type events is shownin the upper right panel of Fig. 3. The majorityof the data from interface events was recorded atdistances ranging from about 30 to 200 km, whilethe intraslab dataset includes ground motionsrecorded at distances Rrup greater than 100 km.The lower panels of Fig. 3 show the distributionof the data for interface and intraslab events byNEHRP site class. As seen from this figure, mostof the strong-motion records available are fromsites classified as NEHRP class C, C/D and D.Only one record from a NEHRP class D/E siteis available and no data was recorded at very softsites (NEHRP class E).

A greater emphasis has been placed on compi-ling and reviewing the available metadata, whichentailed an evaluation of the earthquake-relatedparameters (i.e., magnitude, location and faultmechanism), classification of subduction eventsby type (i.e., interface or intraslab), computationof source-to-site distance metrics and characteri-sation of site conditions at recording stations usingdifferent parameters (i.e., surface geology de-scriptors, shear-wave velocity profiles, natural siteperiod and normalised response spectral shapes).Site classes were assigned to the stations in Peruand Chile following various classification schemesused in ground-motion prediction equations forsubduction-zone environments, such as theNEHRP classification used by Atkinson andBoore (2003, 2008), the New Zealand (NZ) siteclassification scheme used by McVerry et al.(2006) and the Japanese (JP) scheme used byZhao et al. (2006b).

2.1 Earthquake-related parameters

The earthquake-related information was collectedfrom various reporting agencies and publicationsand ranked by preferred importance order. The

epicentral locations and depths of the events usedin this study were selected as follows: specialstudies of mainshock and aftershock sequenceswith accurate relocations, determinations pub-lished in the Centennial catalogue (Engdahl andVillaseñor 2002) and locations and depths deter-mined by the International Seismological Centre(ISC). For the more recent events not included inthe ISC catalogue, the location estimated by theNational Earthquake Information Center (NEIC)was adopted. Regional determinations reportedby local agencies (e.g., Department of Geophysics,University of Chile; IGP) were also used in thisstudy when appropriate. The moment magnitude(MW) estimates and focal mechanism solutionsfor the earthquakes whose data are used in thisstudy were obtained from the Harvard CentroidMoment Tensor database (CMT) when available,which was generally the case for the large earth-quakes included herein with magnitude greaterthan 5.0 which occurred after 1976. For all pre-1976 events, moment magnitudes and focal mech-anism solutions were collected from individualstudies (e.g., Hartzell and Langer 1993; Pachecoand Sykes 2002; Abe 1972). Other instrumen-tal measures of magnitude were collected fromthe online-catalogues of the different reportingagencies (i.e., ISC and NEIC) and were also in-cluded in the metadata. Surface-wave magnitude(MS) and body-wave magnitude (mb) estimates ofthe Peruvian–Chilean earthquakes determined bythe ISC were collected; however, in cases whereISC magnitude determinations were not available,those estimated by NEIC were used instead. Noestimates of moment magnitude (MW) were avail-able for the 5 January 1974 (MS 6.6) event andthe 3 March 1985 (MS 6.4) aftershock. It was,therefore, assumed that MS estimates for theseevents were equivalent to moment magnitude es-timates (MW). This approximation was validatedby plotting moment magnitude values against thedifferent magnitude scales for the events withMW, MS and mb data reported. The characteristicsof the events contributing data to the presentstudy are listed in Table 1; their locations and focalmechanisms are shown in Fig. 2.

Additionally, the earthquake events were clas-sified in terms of the physical processes with whichthey were associated (i.e., interface or intraslab

26 J Seismol (2011) 15:19–41

Tab

le1

Sum

mar

yof

the

eart

hqua

kes

reco

rded

InP

eru

and

Chi

le,w

hose

data

are

used

for

this

stud

y.M

Wes

tim

ates

wer

eob

tain

edfr

omH

arva

rdC

MT

cata

logu

e,ex

cept

for

thos

eev

ents

not

incl

uded

inth

ere,

for

whi

chM

Ses

tim

ates

have

been

liste

din

stea

d(v

alue

sfo

llow

edby

aste

risk

).T

heso

urce

ofth

efa

ult

geom

etry

used

toco

mpu

teth

eru

ptur

edi

stan

ce(R

rup)

isal

solis

ted

alon

gw

ith

the

num

ber

ofre

cord

sav

aila

ble

from

each

even

tand

the

dist

ance

and

PG

Ara

nge.

Oth

erpa

ram

eter

slis

ted

incl

ude

the

hypo

cent

rall

ocat

ion,

the

styl

e-of

-fau

ltin

gan

ddi

men

sion

sof

the

faul

trup

ture

EQ

Eve

ntda

teE

pice

ntre

aD

epth

Mb W

S-of

-Fc

Typ

eF

ault

Fau

ltor

ient

atio

neF

ault

dept

han

dle

ngth

f#

Rru

pP

GA

Dat

a

IDan

dti

me

Lat

Lon

[km

]ge

omet

rySt

rike

[◦]

Dip

[◦]

Hto

pH

hing

e/R

Lre

cord

sra

nge

rang

eav

aila

bilit

y[U

TC

][◦

S][◦

W]

refd

[km

]H

bott

om[k

m]

[km

][c

m/s

2 ][k

m]

117

/10/

1966

10.8

0778

.684

20.7

8.1

RIn

terf

ace

Abe

(197

2)33

512

4.5

3580

116

839

6C

OSM

OSg

[21:

41:5

6]2

31/0

5/19

709.

248

78.8

4073

.08.

0N

Intr

asla

bA

be(1

972)

340

5336

9013

01

265

129

CO

SMO

Sg

[20:

23:3

2]3

05/0

1/19

7412

.360

76.3

9082

.06.

6*N

Intr

asla

bF

rom

scal

ing

315

35*

7687

232

105–

107

88–1

69C

OSM

OSg

[08:

33:5

1]re

lati

ons

403

/10/

1974

12.3

9077

.760

15.0

8.1

RIn

terf

ace

Har

tzel

land

350

11/3

02.

424

/52

250

234

–37

196–

245

CO

SMO

Sg

[14:

21:2

9]L

ange

r(1

993)

509

/11/

1974

12.5

8777

.705

15.0

7.1

RIn

terf

ace

Har

tzel

land

350

114

2160

273

–82

49–1

18C

OSM

OSg

[12:

59:5

2]L

ange

r(1

993)

607

/11/

1981

32.2

3271

.379

63.9

6.9

NIn

tras

lab

Fro

msc

alin

g34

586

5276

343

54–9

328

5–57

1R

EN

AD

ICh

[03:

29:5

2]re

lati

ons

(upo

nre

ques

t)7

03/0

3/19

8533

.115

71.6

1640

.08.

0R

Inte

rfac

eM

endo

za5

15/3

06.

426

/71

255

2526

–215

24–7

07C

OSM

OSg

[22:

47:0

9]et

al.

(199

4)8

03/0

3/19

8532

.829

71.2

1122

.86.

4*–

Inte

rfac

eF

rom

scal

ing

1126

1828

193

31–8

332

–187

CO

SMO

Sg

[23:

38:3

0]re

lati

ons

904

/03/

1985

33.8

3771

.426

38.0

6.3

RIn

terf

ace

Cho

yan

d21

2835

4112

289

–132

57–2

30R

EN

AD

ICh

[15:

01:0

8]D

ewey

(upo

n(1

988)

requ

est)

1025

/03/

1985

34.1

9872

.233

23.0

6.4

RIn

terf

ace

Cho

yan

d10

2021

2513

275

–119

31–1

01R

EN

AD

ICh

[05:

14:3

3]D

ewey

(upo

n(1

988)

requ

est)

J Seismol (2011) 15:19–41 27

1109

/04/

1985

34.0

6071

.589

38.0

7.1

RIn

terf

ace

Cho

yan

d0

2133

4326

936

–197

21–1

58C

OSM

OSg

[01:

57:0

1]D

ewey

(198

8)12

15/1

0/19

9731

.020

71.2

3068

.07.

1N

Intr

asla

bF

rom

scal

ing

173

8054

8243

372

–161

50–3

47R

EN

AD

ICh

[01:

03:3

5]re

lati

ons

(upo

nre

ques

t)13

23/0

6/20

0116

.200

73.7

5029

.08.

4R

Inte

rfac

eP

ritc

hard

310

11/2

57.

525

/70

310

762

–231

31–3

30R

EN

AD

ICh

[20:

33:1

5]et

al.

(upo

n(2

007)

requ

est)

1413

/06/

2005

20.0

1069

.240

108.

07.

8N

Intr

asla

bD

elou

isan

d17

515

/35

105.

411

5/13

211

023

108–

420

18–7

08C

ISM

IDi

[22:

44:3

2]L

egra

ndR

EN

AD

ICh

(200

7)(u

pon

requ

est)

1515

/08/

2007

13.4

9076

.850

18.0

8.0

RIn

terf

ace

Jian

dZ

eng

323

273.

552

190

1337

–139

19–4

88C

ISM

IDi

[23:

40:5

9](2

007)

IGP

j(u

pon

requ

est)

Rre

vers

e,N

norm

al,H

top

dept

hto

top,

Hbo

ttom

dept

hto

bott

om,R

Lru

ptur

ele

ngth

asm

easu

red

alon

gth

est

rike

a Epi

cent

relo

cati

ons

and

dept

hsfr

omth

efo

llow

ing

refe

renc

es:E

ngda

hlan

dV

illas

eñor

(200

2)[E

vent

s1,

2,6,

8],L

ange

ran

dSp

ence

(199

5)[E

vent

s3,

4,5]

,Cho

yan

dD

ewey

(198

8)[E

vent

s7,

9,10

,11]

,[3]

Par

doet

al.(

2002

a)[E

vent

12],

Tav

era

etal

.(20

02)

[Eve

nt13

],D

elou

isan

dL

egra

nd(2

007)

[Eve

nt14

],T

aver

aet

al.(

2008

)[E

vent

15]

bM

wes

tim

ates

are

take

nfr

omth

eC

MT

cata

logu

e.F

orev

ents

earl

ier

than

1976

,the

valu

esre

port

edco

me

from

the

follo

win

gre

fere

nces

:Abe

(197

2)an

dSt

aude

r(1

975)

[Eve

nts

1,2]

,Lan

ger

and

Spen

ce(1

995)

[Eve

nt3]

,Har

tzel

land

Lan

ger

(199

3)[E

vent

s4,

5]c St

yle-

of-F

ault

ing:

R,N

follo

win

gth

eW

ells

and

Cop

pers

mit

h(1

994)

defi

niti

ons

dF

ault

plan

edi

men

sion

sof

even

tsof

unkn

own

geom

etry

have

been

defi

ned

usin

gth

eSt

rass

eret

al.(

2010

)sc

alin

gre

lati

ons

for

subd

ucti

on-z

one

even

tse F

ault

plan

eor

ient

atio

nfr

omth

ese

lect

edfa

ult

mod

el.E

vent

sfo

rw

hich

afi

nite

sour

cem

odel

isno

tav

aila

ble,

the

stri

kean

ddi

pha

vebe

ense

lect

edfr

omth

etw

ose

tsof

angl

esre

port

edin

the

Har

vard

CM

Tca

talo

gue

(see

text

for

expl

anat

ion)

.Tw

odi

pva

lues

are

repo

rted

for

hing

edfa

ultm

odel

sf F

orm

ulti

-seg

men

tmod

els,

the

dept

hto

the

hing

ein

the

faul

tis

also

repo

rted

(Hhi

nge)

g Stro

ng-m

otio

nre

cord

sdo

wnl

oada

ble

atht

tp://

db.c

osm

os-e

q.or

ghIn

tere

sted

read

ers

shal

lcon

tact

the

4thau

thor

rega

rdin

gst

rong

-mot

ion

reco

rds

avai

labi

lity

(bor

osch

@ce

c.uc

hile

.cl)

i Stro

ng-m

otio

nre

cord

sdo

wnl

oada

ble

atht

tp://

ww

w.c

ism

id-u

ni.o

rg/

j Inte

rest

edre

ader

ssh

allc

onta

ctth

e6th

auth

orre

gard

ing

stro

ng-m

otio

nre

cord

sav

aila

bilit

y(h

erna

ndo.

tave

ra@

igp.

gob.

pe)

28 J Seismol (2011) 15:19–41

activity). The earthquake classification was madeon the basis of focal mechanism, epicentral lo-cation, depth and relative position with respectto the trench axis. The dominant mechanism ofinterface-type earthquakes corresponds to thrustfaulting on shallow-dipping planes that are ori-ented approximately parallel to the trench axis. Atdepths greater than the coupled plate interface,the stress regime changes from compressional totensional and thus normal faulting prevails. Thesenormal mechanism events are associated with in-traslab activity occurring within the subductedNazca slab, at some distance down-dip fromthe seismically coupled interface. At intermedi-ate depths, two types of intraslab earthquakeshave been identified in this region (e.g., Lemoineet al. 2002): slab-push and slab-pull, which areassociated with down-plate compression and ex-tension, respectively. Along the Peru–Chile sub-duction zone, the occurrence of slab-pull eventsis relatively common in comparison with slab-push events, although some slab-push events haveoccurred in North-Central Chile (the 15 October1997 Punitaqui earthquake) and Central Peru (the5 and 29 April 1991 earthquakes).

The differentiation between interface and in-traslab events was performed using the definitionsof style-of-faulting of Wells and Coppersmith(1994) and the event depth and location with re-spect to the trench axis. As seen in Table 1, theinterface events in this catalogue have a reversemechanism and are limited to a maximum depthof 40 km, which is consistent with the maxi-mum depth extent of the seismically coupled zonefound along different segments of the Peru–Chilesubduction zone (e.g., Comte et al. 1994; Comteand Suárez 1995; Tichelaar and Ruff 1991). Onthe other hand, intraslab-type events in Table 1have a normal mechanism and occur within theNazca slab at depths from about 60 to 110 km.The location of these events in 3D space combinedwith information regarding their focal mechanism(Fig. 1 and lower left panel of Fig. 2) allows a fairlyunambiguous classification between interface andintraslab events, particularly since the geometryof the subducting Nazca slab has been extensivelydocumented. The geometry of the subductingNazca plate is characterised by variations in thedip angle along the strike of the trench (Barazangi

and Isacks 1976; Jordan et al. 1983; Cahill andIsacks 1992). Between latitude 8◦ and 45◦ S, thesubducting Nazca plate is divided into four seg-ments: northern and central Peru, from 8◦ to 15◦ S,where the subducted Nazca plate has a shallow dipof about 10◦; southern Peru and northern Chile,from 15◦ to 27◦ S, where the Nazca Plate descendswith a dip of 25◦ to 30◦. In Central Chile, from27◦ to 33◦ S, the slab is again relatively flat, witha shallow dip angle of about 10◦, and in southernChile, from 33◦ to 45◦ S, the dip of the subductedslab increases to 30◦.

The depth extent of the seismically coupledplate interface along the Peru–Chile subductionzone is similarly well-documented. It has beenestimated from the maximum depth of shallow-dipping reverse events (e.g., Tichelaar and Ruff1991; Suárez and Comte 1993; Comte et al. 1994;Comte and Suárez 1995) and from the depth tran-sition from compressional to extensional stressregime (e.g., Comte and Suárez 1995; Pardo et al.2002b). Based on the maximum depth of large(MW > 6) under-thrusting events located teleseis-mically, Tichelaar and Ruff (1991) suggested thatthe maximum depth of the seismically coupledzone between plates along Chile extends downto 48–53 km and that there is a change in themaximum depth north of latitude 28◦ S, where thecoupled zone extends to depths of 36–41 km. Incontrast, studies using both locally and teleseis-mically recorded data in Northern Chile (Comteet al. 1994; Comte and Suárez 1995) suggestthat the coupling zone, as defined by the maxi-mum depth observed for shallow-dipping reverseevents, extends consistently to about 40 ± 10 kmand no variations along the strike of the trenchare appreciable. The maximum depth of the cou-pling zone may, however, extend up to 60 ±10 km, if the depth transition from compres-sional to tensional stress regime observed alongthe upper part of the subducting slab is consid-ered (Comte and Suárez 1995). This transitionof stress field along the Northern Chile segmentoccurs at depths greater than the maximum depthat which shallow-dipping reverse events are ob-served (∼40 km). Along the Central Chile seg-ment of the subduction zone, the maximum depthof the plate interface has also been estimatedto be about 60 km (Pardo et al. 2002a), which

J Seismol (2011) 15:19–41 29

is in agreement with the aforementioned studiesalong different segments of the Chilean subduc-tion zone.

2.2 Station information and assignmentof site conditions

The coordinates of the stations used in this study,type of instrument and instrument housing arelisted in Tables 2 and 3 and their geographicaldistribution is shown in Fig. 2. For the stations inCentral Chile that recorded the 1985 Valparaisoearthquake sequence, some of which are no longerin operation, the station coordinates listed corre-spond to those reported by Campbell et al. (1989,1990), which were validated against satellite im-agery (i.e., Google Earth) to ensure accuracy. In-formation on the type and location of instrument(i.e., type of building) was also obtained fromthese references, from the accelerogram headingsand from the websites of the network operators(IGP, CISMID, RENADIC and DGF-DIC). In-formation on the majority of the Peruvian stationsincluded herein has already been presented inTavera et al. (2008).

Tables 2 and 3 also summarise all geologicaland geotechnical information collected for thesites contributing data to this study. Site condi-tions assigned to the stations in Central Chilewere based on information collected from a num-ber of references including descriptions of thesurface geology (EERI 1986; Çelebi 1987, 1988;Campbell et al. 1989, 1990, Midorikawa et al. 1991;Midorikawa 1992), the site categories follow-ing the Chilean seismic design code assigned byRiddell (1995) and NEHRP site classes assignedby Atkinson and Boore (2003) to the Chileansites whose data were included in the regressiondatabase for subduction-zone events. Shear-wavevelocity (VS) profiles obtained by Araneda andSaragoni (1994), Midorikawa et al. (1991) andMidorikawa (1992) in addition to the natural pe-riod of the Chilean sites determined by Luppichini(2004) using the records of the 1985 Valparaisoearthquake, were also used. Site conditions at thestrong-motion stations in Northern Chile are stillunder investigation and geological and geotech-nical information for a number of these stations

has not yet been made available to the widerengineering community. It is believed, however,that recording sites in Northern Chile can be clas-sified as NEHRP class C, with an average shear-wave velocity over the top 30 m, VS30, between400 and 600 m/s (Boroschek and Comte 2006).Therefore, site conditions assigned to these siteswere only based on information from descriptionsof the local geology (SNGM 1982), VS profilesobtained from spectral analysis of surface waves(SASW) measurements at the stations in Aricaand Poconchile (Cortez-Flores 2004), and naturalsite periods estimated by site response analysis forthe Arica and Poconchile stations (Cortez-Flores2004).

Site conditions assigned to the stations inCentral and Southern Peru were based on de-scriptions of the surface geology (EERI 2007;Bernal and Tavera 2007a, b) and the site cate-gory (i.e., rock, soil or firm ground) assigned byRodriguez-Marek et al. (2007). Shear-wave veloc-ity (VS) profiles obtained from SASW measure-ments at the stations in Ica (Rosenblad and Bay2008) and the stations in Moquegua and Tacna(Cortez-Flores 2004), as well as the VS profilesestimate by Bernal and Tavera (2007a, b) usingan infinite flat-layered half-space model were alsoused. Additionally, the natural site period as in-terpreted from the microzonation map of Lima(Aguilar Bardales and Alva Hurtado 2007) andthat estimated by site response analysis for theMoquegua and Tacna sites (Cortez-Flores 2004)were included. Information on the site conditionsof the majority of the Peruvian stations includedherein has been also been reported in Tavera et al.(2008).

Besides the site information collected, the spec-tral shapes of the records were considered bynormalising the response spectra by their PGAvalue (for all records) and by dividing the spectrarecorded at soil stations by the spectrum obtainedon rock, for stations sufficiently close to one an-other. The natural period for each site computedfrom earthquake records, T0,REC, was also esti-mated and used as a guide for the assignationof site classes, following the empirical site clas-sification approach (JP) adopted by Zhao et al.(2006a) which defines the site period as that cor-responding to the highest H/V response spectral

30 J Seismol (2011) 15:19–41

Tab

le2

Sum

mar

yof

char

acte

rist

ics

ofC

hile

anst

rong

-mot

ion

stat

ions

used

inth

isst

udy

Inst

rum

ent&

stat

ion

info

rmat

ion

Geo

logi

cal&

geot

echn

ical

info

rmat

ion

Site

clas

ses

assi

gned

Cod

eaN

ame

Lat

[◦S]

Lon

[◦W

]IT

bIL

cSu

rfac

ege

olog

ydSC

e AB

SCf C

FSC

g RV

h S30

Ti 0,

LU

PT

j 0,C

FT

k 0,R

EC

NH

lN

Zm

JPn

CO

o

AC

AA

rica

-Cas

a18

.482

70.3

08S

B1

Mar

ine

and

–C

2–

432[

1]–

TS1

=0.

150.

13–0

.34

CC

IIII

[RE

N]

cont

inen

tal

TS2

=0.

19se

dim

ents

onro

ck[1]

;Se

dim

ents

[2]A

CO

Ari

ca18

.474

70.3

13S

BM

arin

ean

d–

C2

–38

9[1]

–T

S1=

0.32

0.33

–0.3

9C

CII

II[R

EN

]C

osta

nera

cont

inen

tal

TS2

=0.

36se

dim

ents

onro

ck[1]

;Se

dim

ents

[2]A

RIE

Ari

ca18

.494

70.3

12E

B1

Vol

cani

cro

ck[1]

;–

B–

1132

[1]–

TS1

=0.

110.

38B

BI

I[D

-D]

Esc

uela

Roc

k[2]

CA

LA

Cal

ama

22.4

5968

.930

EB

Sedi

men

ts[2]

;–

––

––

–∼0

.10

CC

IIII

[D-D

]H

ospi

tal

Dee

pse

dim

ents

[3]C

AU

Cau

quen

es35

.97

72.3

2S

B2

Allu

vium

[4,5] ;

D–

II64

8[2]

0.45

–0.

40–0

.62

C/D

CII

III

[RE

N]

Den

segr

avel

[6]C

HIL

Chi

llán-

Vie

jo36

.60

72.1

0S

B2

Allu

vium

[4];

––

II56

8[2]

0.77

–0.

35–0

.56

C/D

CII

III

[RE

N]

Soft

allu

vium

[5];

Den

segr

avel

[6]C

ON

SC

onst

ituc

ión

35.3

372

.41

SB

2G

rani

te[4,

7] ;D

–II

I59

5[2]

0.83

–∼0

.74

C/D

DII

III

I[R

EN

]P

aleo

zoic

intr

usiv

e[5] ;

Med

ium

dens

ity

sand

[6]C

UY

Cuy

a19

.160

70.1

77S

USe

dim

enta

ryro

ck–

––

––

–0.

22–0

.33

CC

IIII

[RE

N]

and

mar

ine

sedi

men

ts[1]

EN

DSa

ntia

go33

.45

70.6

5P

B6

Fir

mgr

avel

[4];

––

–51

3[3]

0.33

–0.

70–0

.81

CC

III

II[R

EN

]E

ndes

aA

lluvi

um[5]

;Sh

allo

wfi

llon

dens

egr

avel

[8]

J Seismol (2011) 15:19–41 31

HU

AH

uala

ñe34

.97

71.8

2S

B1

Allu

vium

[4,7] ;

B–

II52

7[2]

0.38

–∼0

.36

C/D

CII

II[R

EN

]D

ense

grav

el[6]

ILL

AIl

lape

l31

.63

71.1

7S

B1

Allu

vium

[4,7] ;

E–

II61

3[2]

0.25

–0.

16–0

.22

CC

IIII

[RE

N]

Den

segr

avel

[6]IL

OIl

oca

34.9

372

.18

SB

1A

lluvi

um[4,

5,7] ;

D–

II55

5[2]

0.33

–0.

22–0

.43

C/D

CII

II[R

EN

]Sa

nd[6]

IQU

Iqui

que-

Idie

m20

.215

70.1

40S

BSe

dim

ents

[2]–

––

––

–∼0

.50

CB

III

II[R

EN

]IQ

UC

Iqui

que-

Inp

20.2

1770

.149

SB

Sedi

men

ts[2]

––

––

––

∼0.5

3C

CII

III

[RE

N]

IQU

IIq

uiqu

e20

.214

70.1

38E

BR

ock[

2] ;R

ock[

3]–

––

––

–∼0

.40

BC

II

[D-D

]H

ospi

tal

ISID

San

Isid

ro32

.90

71.2

7S

UA

lluvi

um[1]

D–

–78

9[2]

0.33

–0.

37C

CII

II[R

EN

]L

IGL

aL

igua

32.4

571

.25

SB

1A

lluvi

um[4,

7] ;D

–II

620[

2]0.

29–

–C

CII

II[R

EN

]D

ense

grav

el[6]

LL

AY

Lla

yL

lay

32.8

470

.97

SB

1So

ftal

luvi

um[4,

7] ;E

–II

610[

2]0.

67–

∼1.

0D

DII

III

I[R

EN

]G

rave

land

soft

lime[

6]L

LO

Llo

lleo

33.5

871

.61

SB

1Sa

ndst

one

and

––

II30

5[2]

0.53

–0.

42–0

.52

C/D

CII

III

[RE

N]

volc

anic

rock

[4,7] ;

Den

sesa

nd[6]

LO

AE

lLoa

22.6

3668

.152

SU

Vol

cani

cro

ck[1]

––

––

––

∼0.1

2B

CI

I[R

EN

]M

EJI

Mej

illon

es-

23.1

0370

.446

EB

Sedi

men

ts[2]

;–

––

––

–0.

33–0

.85

CD

III

III

[D-D

]H

ospi

tal

very

deep

sand

s[3]

ME

LP

Mel

ipill

a33

.68

71.2

2S

B1

Allu

vium

[4];

C–

II72

4[2]

0.30

–0.

20–0

.35

CC

IIII

[RE

N]

Den

sesa

nd[6]

;G

rani

te[7]

PA

PP

apud

o32

.51

71.4

5S

B1

Gra

nite

[4,7] ;

B–

I51

7[2]

0.34

–0.

26–0

.36

C/D

CII

II[R

EN

]W

eath

ered

rock

[6]P

ICA

Pic

a–

20.4

9269

.330

EB

Sedi

men

ts[2]

––

––

––

∼0.3

5C

CII

II[D

-D]

Hos

pita

l

32 J Seismol (2011) 15:19–41

Tab

le2

(con

tinu

ed)

Inst

rum

ent&

stat

ion

info

rmat

ion

Geo

logi

cal&

geot

echn

ical

info

rmat

ion

Site

clas

ses

assi

gned

Cod

eaN

ame

Lat

[◦S]

Lon

[◦W

]IT

bIL

cSu

rfac

ege

olog

ydSC

e AB

SCf C

FSC

g RV

h S30

Ti 0,

LU

PT

j 0,C

FT

k 0,R

EC

NH

lN

Zm

JPn

CO

o

PIC

HP

ichi

lem

u34

.38

72.0

2S

B1

Slat

es,s

ands

tone

,B

–I

623[

2]0.

33–

∼0.2

3C

BII

I[R

EN

]lim

esto

ne[4,

7] ;R

ock[

6]P

ISP

isag

ua19

.595

70.2

12S

USh

allo

wfi

llon

––

––

––

0.10

–0.3

3C

BII

I[R

EN

]w

eath

ered

rock

[3]P

OC

O1

Poc

onch

ile1

18.4

5670

.067

SB

Mar

ine

and

–C

2–

511[

1]–

TS1

=0.

240.

21–0

.57

CC

IIII

[RE

N]

cont

inen

tal

TS2

=0.

22se

dim

ents

onro

ck[1]

PO

CO

2P

ocon

chile

218

.457

70.1

07E

BM

arin

ean

d–

C2

––

–T

S1=

0.24

0.21

CC

IIII

[D-D

]co

ntin

enta

lT

S2=

0.22

sedi

men

tson

rock

[1]P

UP

utre

18.1

9769

.574

SU

Wea

ther

edro

ck[3]

––

––

––

0.38

–0.5

6C

BII

III

[RE

N]

QU

INQ

uint

ay33

.20

71.6

8S

SP

aleo

zoic

––

I59

5[2]

0.50

-0.

48–0

.66

CB

III

[RE

N]

intr

usiv

es[5]

;R

ock[

6]R

AP

Rap

el34

.03

71.5

8R

TSe

dim

ents

[4,7] ;

B–

I30

10[2]

0.40

–0.

10–0

.29

AA

II

[RE

N]

Pal

eozo

icin

trus

ives

[[5] ;

Roc

k[6]

SAN

TSa

ntia

go33

.47

70.6

7S

B3

Fir

mgr

avel

[4];

––

––

0.65

–0.

66–0

.91

CC

III

II[R

EN

]A

lluvi

um[5]

SFD

OSa

nF

erna

ndo

34.6

071

.00

SB

1A

lluvi

um[4,

7] ;D

–II

543[

2]0.

36–

0.22

–0.4

6C

CII

II[R

EN

]D

ense

grav

el[6]

SFE

LSa

nF

elip

e32

.75

70.7

3S

B1

Allu

vium

[4,7] ;

D–

II50

2[2]

0.50

–∼0

.12

CC

IIII

[RE

N]

Den

segr

avel

[6]T

AL

Tal

ca35

.43

71.6

7S

B1

Allu

vium

[4,7] ;

E–

II59

8[2]

0.83

–0.

17C

CII

III

[RE

N]

Den

segr

avel

[6]T

CP

Toc

opill

a22

.104

70.2

14E

UR

ock[

2]–

––

––

–∼0

.10

BB

II

[D-D

]

J Seismol (2011) 15:19–41 33

UT

FSM

Val

para

iso

33.0

371

.60

SB

1V

olca

nic

rock

[4,7] ;

B–

I14

21[2]

1.00

–0.

78–0

.87

AA

II

[RE

N]

UT

FSM

Roc

k[6]

VA

LM

DV

alpa

rais

o33

.03

71.6

4S

RF

ill[4,

6] ;So

il[7] ;

D–

III

360[

3]0.

67–

–D

DII

III

I[R

EN

]E

lAlm

endr

alA

rtif

icia

lfill

[8]V

EN

TV

enta

nas

33.0

371

.62

SB

6L

oose

sand

[4];

D–

III

331[

2]0.

67–

0.76

–1.0

DD

III

III

[RE

N]

Allu

vium

[5];

Sand

[6,7]

VIL

Los

Vilo

s31

.92

71.5

0S

B1

Sedi

men

tary

B–

I12

15[2]

––

0.26

AB

II

[RE

N]

rock

[4,7] ;

Roc

k[6]

VM

AR

Viñ

ade

lMar

33.0

271

.55

SB

10A

lluvi

uman

d–

–II

I27

3[3]

0.50

–0.

50–0

.80

DD

III

I[R

EN

]sa

nd[4]

;San

d[6]

ZA

PZ

apal

lar

32.5

571

.46

SB

1R

ock[

6] ;G

rani

te[7]

B–

I60

5[2]

0.41

–∼0

.18

CB

IIII

[RE

N]

a Stat

ion

code

,fol

low

edby

the

netw

ork:

RE

N=R

EN

AD

IC;D

-D=D

GC

-DIC

bIn

stru

men

ttyp

e:E

=ET

NA

;P=P

K-1

30;R

=RF

T-2

50;S

=SM

A-1

;c In

stru

men

tloc

atio

n:B

=bui

ldin

g,fo

llow

edby

num

ber

ofst

orey

sif

know

n;S=

shel

ter;

T=t

unne

l;U

=unk

now

ndD

escr

ipti

onof

the

surf

ace

geol

ogy,

base

don

the

follo

win

gre

fere

nces

:[1]

Geo

logi

cm

apof

Chi

le(S

NG

M19

82);

[2]

Alv

aH

urta

do(2

005)

;[3]

this

stud

y;[4

]Ç

eleb

i(1

988)

;[5]

Cam

pbel

leta

l.(1

990)

;[6]

Rid

dell

(199

5);[

7]E

ER

Ire

conn

aiss

ance

repo

rt(1

986)

;[8]

Mid

orik

awa

etal

.(19

91)

and

Mid

orik

awa

(199

2)e N

EH

RP

site

clas

ses

assi

gned

byA

tkin

son

and

Boo

re(2

003,

2008

)f Si

tecl

asse

sas

sign

edby

Cor

tez-

Flo

res

(200

4)fo

llow

ing

the

Rod

rigu

ez-M

arek

etal

.(20

01)

site

clas

sifi

cati

onsc

hem

e:B

=Roc

k;C

2=Sh

allo

wst

iffs

oil

g Soil

clas

ses

assi

gned

byR

idde

ll(1

995)

follo

win

gth

e19

93C

hile

anse

ism

icde

sign

code

prov

isio

nshA

vera

gesh

ear-

wav

eve

loci

tyov

erth

eto

p30

m,i

nm

/s,d

eter

min

edfr

om:[

1]V

spr

ofile

sob

tain

edby

Cor

tez-

Flo

res

(200

4)us

ing

SASW

;[2]

VS

prof

iles

dete

rmin

edby

Ara

neda

and

Sara

goni

(199

4);V

Spr

ofile

sob

tain

edby

Mid

orik

awa

etal

.(19

91)

and

Mid

orik

awa

(199

2).F

orca

lcul

atio

npu

rpos

es,w

hen

VS

data

wer

eav

aila

ble

atde

pths

<30

m,t

heV

Sva

lue

ofth

ela

stla

yer

was

assu

med

cons

tant

to30

mde

pth

i Pre

dom

inan

tsit

epe

riod

,in

seco

nds,

dete

rmin

edby

Lup

pich

ini(

2004

).V

alue

slis

ted

asre

port

edby

Rui

zan

dSa

rago

ni(2

005)

j Pre

dom

inan

tsi

tepe

riod

dete

rmin

edby

Cor

tez-

Flo

res

(200

4).T

S1w

ases

tim

ated

asth

epe

riod

corr

espo

ndin

gto

the

max

imum

rati

oof

resp

onse

spec

tra

atth

esu

rfac

eov

erth

ere

spon

sesp

ectr

aof

outc

rop

inpu

tmot

ion

and

TS2

corr

espo

nds

toth

ech

arac

teri

stic

site

peri

odca

lcul

ated

from

the

equa

tion

TS

=4H

/V

SkP

redo

min

ants

ite

peri

odca

lcul

ated

from

acce

lero

gram

byco

nsid

erin

gth

eH

/Vra

tio

ofth

ere

spon

sesp

ectr

a,fo

llow

ing

the

appr

oach

ofZ

hao

etal

.(20

06a)

.The

low

eran

dup

per

boun

dari

esof

the

inte

rval

repo

rted

corr

espo

ndto

the

max

imum

and

min

imum

valu

esof

the

natu

rals

ite

peri

odfo

und

whe

nus

ing

mul

tipl

ere

cord

sfr

omth

esa

me

stat

ion.

Val

ues

are

only

liste

dfo

rth

ose

reco

rds

who

seve

rtic

alco

mpo

nent

isav

aila

ble

l Site

clas

sac

cord

ing

toth

eN

EH

RP

(199

7)pr

ovis

ions

used

inan

alys

esm

Site

clas

sas

sign

edfo

llow

ing

the

New

Zea

land

site

clas

sifi

cati

on,w

hich

isba

sed

onsu

rfac

ege

olog

y,ge

otec

hnic

alpr

oper

ties

,nat

ural

site

peri

odan

dde

pth

tobe

droc

k(s

eeM

cVer

ryet

al.2

006

for

deta

ils)

nSi

tecl

ass

assi

gned

follo

win

gth

eZ

hao

etal

.(20

06b)

sche

me,

cons

ider

ing

the

natu

ralp

erio

dof

the

site

oSi

tecl

ass

assi

gned

toco

mpu

teth

ede

sign

load

spr

escr

ibed

byth

e19

96C

hile

anse

ism

icco

de

34 J Seismol (2011) 15:19–41

Tab

le3

Info

rmat

ion

onP

eruv

ian

stro

ng-m

otio

nst

atio

nsus

edin

this

stud

y

Cod

eN

ame

ITL

at[◦

S]L

on[◦

W]

Loc

Surf

ace

geol

ogya

SCb R

MV

c S30

Td 0,

CIS

Te 0,

RE

CN

Hf

NZ

gJP

hC

Oi

[m/s

]

AN

C[I

GP

]A

ncon

D11

.776

77.1

50U

Allu

vial

grav

el(s

oil)

[2]S

280[

5]0.

2–0.

30.

300.

10C

/DC

IIII

AN

R[C

ER

]A

sam

blea

D12

.123

76.9

76B

Allu

vial

grav

el(s

oil)

[2]F

G20

5[5]

0.2–

0.3

0.50

0.15

DC

IIII

Nac

iona

lde

Rec

tore

sC

AL

[CIS

]C

alla

oE

12.0

6077

.150

SSo

ftso

il[1] ;

Soft

clay

[2];

S75

[5]0.

5–0.

60.

530.

52D

/EE

IVII

IG

ranu

lar

fill

over

fine

stra

tifi

edso

ils[3]

CD

L-C

IP[C

IS]

CD

L-C

IPE

12.0

9277

.049

SD

ense

,sti

ffgr

avel

depo

sit

FG

0.1–

0.2

0.82

0.30

DC

III

II(L

ima

Con

glom

erat

e)[1]

;A

lluvi

algr

avel

(soi

l)2]

CE

R[C

ER

]C

eres

isE

12.1

0376

.998

UA

lluvi

algr

avel

(soi

l)[2]

FG

0.1–

0.2

0.28

0.45

DC

III

IIC

SM[C

IS]

Cis

mid

D12

.013

77.0

50B

1D

ense

,sti

ffgr

avel

depo

sit

FG

184[

5]0.

2–0.

30.

050.

10C

CII

I(L

ima

Con

glom

erat

e)[1]

;A

lluvi

algr

avel

(soi

l)[2]

GE

O[I

GP

]G

eolo

gica

lA

12.0

876

.95

UC

oars

ede

nse

grav

el–

CB

IIII

Inst

itut

eH

UA

[IG

P]

Cas

aH

uaco

–A

12.1

376

.98

UA

lluvi

alde

posi

ts–

CB

IIII

Las

Gar

deni

asIC

A2

[CIS

]Ic

a2

A14

.089

75.7

32B

Silt

ysa

nd,s

oil[1

]S

312

–0.

720.

48D

CII

III

LM

OL

[IG

P]

La

Mol

ina

A12

.085

76.9

48U

Allu

vial

depo

sits

Uni

vers

idad

(sof

tcla

ysan

dsa

nd)[4

]A

grar

iaM

AY

[IG

P]

May

oraz

goD

12.0

5576

.944

USa

ndan

dsi

lt[2]

S27

6[5]

0.2–

0.3

0.22

0.20

CC

III

MO

L[C

IS]

Mol

ina

E12

.10

76.8

9B

Shal

low

soil

R38

0[5]

0.2–

0.4

0.13

0.20

CC

III

over

lyin

gde

nse

Lim

aC

ongl

omer

ate;

Sand

[2]

J Seismol (2011) 15:19–41 35

MO

Q1

[CIS

]M

oque

ga1

A17

.187

70.9

29S

Allu

vial

depo

sits

573

0.11

–0.1

8C

BII

II(s

andy

grav

els)

[5]N

NA

[IG

P]

Ñañ

aD

11.9

8776

.389

UR

ock[

2]R

–0.

100.

22B

BI

IP

CN

[IG

P]

Par

cona

D14

.042

75.6

99U

Soil[

1]S

456

0.42

0.54

C/D

CII

III

PU

CP

[PU

CP

]U

nive

rsid

adD

12.0

7477

.080

BA

lluvi

algr

avel

(soi

l)[2]

FG

125[

5]0.

2–0.

30.

900.

90D

DII

III

Cat

olic

ade

lPer

uR

IN[C

ER

]R

inco

nada

D12

.084

76.9

21U

Fill

cons

isti

ngof

sand

,S

200[

5]0.

2–0.

30.

320.

30C

/DC

IIII

silt

and

grav

el[2]

a Des

crip

tion

ofsu

rfac

ege

olog

ypr

ofile

,bas

edon

the

follo

win

gre

fere

nces

:[1]

EE

RI

(200

7)[2

]B

erna

land

Tav

era

(200

7a,b

)[3

]in

form

atio

npr

ovid

edby

the

stro

ng-

mot

ion

netw

ork

inth

eac

cele

rogr

amhe

adin

g[4

]Esp

inos

aet

al.(

1977

)[5

]Cor

tez-

Flo

res

(200

4)bSi

tecl

ass

assi

gned

byR

odri

guez

-Mar

eket

al.(

2007

)c A

vera

gesh

ear-

wav

eve

loci

tyov

erth

eto

p30

m.F

orth

eIc

ast

atio

ns,t

his

isba

sed

onth

eV

Spr

ofile

sob

tain

edby

Ros

enbl

adan

dB

ay(2

008)

usin

gSA

SW.F

orth

eL

ima

stat

ions

,the

valu

eta

bula

ted

isa

tent

ativ

ees

tim

ate

ofV

S30

base

don

the

VS

prof

ilein

ferr

edby

Ber

nala

ndT

aver

a(2

007a

,b)

usin

gan

infi

nite

flat

-lay

ered

half

-spa

cem

odel

dN

atur

alsi

tepe

riod

(T0)

infe

rred

from

the

mic

rozo

nati

onm

apof

Lim

a(A

guila

rB

arda

les

and

Alv

aH

urta

do20

07).

Val

ues

are

not

avai

labl

efo

rth

eN

NA

stat

ion

inL

ima,

nor

for

the

Ica

stat

ions

e Pre

dom

inan

tpe

riod

calc

ulat

edfr

omac

cele

rogr

amby

cons

ider

ing

the

H/V

rati

oof

the

resp

onse

spec

tra,

follo

win

gth

eap

proa

chof

Zha

oet

al.(

2006

a).T

heto

pva

lue

corr

espo

nds

toth

eea

st-w

estc

ompo

nent

ofm

otio

n,w

hile

the

bott

omva

lue

corr

espo

nds

toth

eno

rth-

sout

hco

mpo

nent

f Site

clas

sac

cord

ing

toth

eN

EH

RP

(199

7)pr

ovis

ions

.T

henu

mbe

rin

brac

kets

corr

espo

nds

toth

eV

S30

valu

eas

sum

edw

hen

expl

icit

lyre

quir

ed,

follo

win

gth

ere

com

men

dati

ons

ofA

tkin

son

and

Boo

re(2

003)

g Site

clas

sac

cord

ing

toth

eN

ewZ

eala

ndsi

tecl

assi

fica

tion

,whi

chis

base

don

surf

ace

geol

ogy,

geot

echn

ical

prop

erti

esan

dde

pth

tobe

droc

k.Se

eM

cVer

ryet

al.(

2006

)fo

rde

tails

hSi

tecl

ass

acco

rdin

gto

the

Zha

oet

al.(

2006

b)sc

hem

e,co

nsid

erin

gV

S30

and

the

natu

ralp

erio

dof

the

site

i Site

clas

sas

sum

edto

com

pute

the

desi

gnlo

ads

pres

crib

edby

the

1977

and

2003

Per

uvia

nse

ism

icco

des

36 J Seismol (2011) 15:19–41

ratio. Tables 2 and 3 also list the site classes as-signed to the different stations following severalclassification schemes: the NEHRP site clas-sification, which is based on the average shear-wave velocity over the top 30 m; the New Zealandclassification scheme used by McVerry et al.(2006), which classifies sites on the basis of the sur-face geology, geotechnical properties, site periodand depth to bedrock; and the JP scheme used byZhao et al. (2006b), which uses the predominantsite period from H/V response spectral ratios.

Due to the inherent limitations of some of thesite data collected, the same level of priority wasnot given to all the various pieces of informationin the assignment of site classes. For instance,only VS profiles determined from measurementsof shear-wave velocity conducted in the field havebeen used for the direct assignment of site classes,and profiles reported from inversions (e.g., Bernaland Tavera 2007a, b) have only been used todistinguish between shallow and deep soil sitessince in several instances these profiles have beenfound to be biased towards low values, leading tosite classifications that are inconsistent with othergeological and geotechnical descriptions. Simi-larly, the VS30 values listed in Table 2, calculatedfrom the VS profiles estimated by Araneda andSaragoni (1994) at a number of sites in centralChile (i.e., LLAY, MEL, ISI), were found to bebiased towards high values. As no information asto the manner in which the VS values published inAraneda and Saragoni (1994) were obtained (i.e.,in situ measurements or numerical modelling),these VS profiles have only been used to iden-tify different soil depths. In addition, the naturalperiod (T0, CIS) derived using ambient noise mea-surements mapped in the microzonation map ofLima (Aguilar Bardales and Alva Hurtado 2007),was generally the preferred input for assigningthe JP site classes to the Peruvian sites as thepredominant period calculated directly from therecords (T0,REC) could be biased due to non-linearity effects. In some instances, however, itwas found that mapped period was inconsistentwith other site descriptors, possibly due to limita-tions of the mapping resolution. For the stations inChile, natural periods estimated by site responseanalysis (Cortez-Flores 2004) were the preferredinput.

Most of the stations in Central Chile are situ-ated on dense alluvial gravel and sand classifiedas NEHRP class C, C/D and D. There are nostations situated on soft soil (NEHRP E); how-ever, the VMAR and V-ALM stations are on deepsand and artificial fill, respectively, and, therefore,exhibit features consistent with soils of mediumdensity. These sites are classified as NEHRP D inview of the large values of the VS30 reported. Onlythree stations are located on hard rock and rock(NEHRP class A and B), and three sites are onsoft/weathered rock classified as NEHRP class C:the RAP, VIL and UTFSM stations are locatedon rock (NEHRP site class B) and ZAP, QUI,PIC sites are on soft/weathered rock (NEHRP siteclass C). Stations in Northern Chile are situatedon volcanic rock and shallow fill on weatheredrock, classified as NEHRP B and C, respectively.The most recent material in this region consistsof Quaternary alluvial and fluvial deposits andmany of the stations are located on such material.These sites are, therefore, classified as NEHRP Cby virtue of the VS30 values estimated for someof those sites (ACA, ACO, POCO1, POCO2) aswell as the shape of the normalised spectra (IQU,MEJI, PICA, CUY). The majority of the stationsin Peru used in this study are situated on allu-vial gravel, sand and silt and have been classifiedas NEHRP class C and D and only one station(NNA) is situated on rock classified as NEHRPclass B. Conversely, the station CAL is locatedclose to the coast in an area of reclaimed landover soft soil, and has been classified as NEHRPclass D/E. Another station located on reclaimedland is RIN, which is located on loose granular fillcomposed of gravel, silt and fine sand.

2.3 Record information

Processing was performed using TSPP: Collec-tion of FORTRAN programs for processing andmanipulating time series, developed by Dr DavidBoore from the United States Geological Sur-vey (Boore 2008). The ground-motion recordingswere reformatted and converted into SMC-formatfiles (see http://nsmp.wr.usgs.gov/smcfmt.html fordetails). When necessary, unevenly sampled datawere interpolated and resampled at 200 samplesper second. Before the application of any process-

J Seismol (2011) 15:19–41 37

ing procedure, non-standard noise (i.e., spuriousspikes) encountered in digitized records from ana-logue instruments (Douglas 2003) was identifiedby visual inspection of the jerk (derivative ofthe acceleration trace). Spikes identified as erro-neous, were removed by replacing the accelera-tion ordinate of the spike with the mean of thepreceding and proceeding accelerations values.For some of the analogue recordings includedin this database, instrument correction has beenalready applied by the data provider and thuswas not applied here. Instrument corrections werenot applied to the remaining records from ana-logue instruments in this database as, in somecases, complete information on the instrumentsresponse was not available and additionally, theapplication of an instrument correction can re-sult in amplification of high-frequency noise intro-duced during the digitization process (Boore andBommer 2005). As a result of the dynamic rangeof the digital instruments (natural frequencies of100 Hz or higher) corrections for instrument char-acteristics were not applied to the digital record-ings included in the database.

The records were processed in a consistentmanner, with individual components individuallyfiltered. Before performing the actual filteringof the record, an initial baseline correction wasapplied to the raw accelerogram (zeroth-ordercorrection). The mean determined from the pre-event portion of the record or the mean computedfrom the whole record if the pre-event portionwas not available was subtracted from the entireacceleration time series. After making this ini-tial baseline correction, the acceleration traceswere integrated without filtering, to check forlong-period drifts that could indicate the pres-ence of offsets in the reference baseline. Inmost cases, baseline offsets were small and thelong-period noise was removed by filtering. Therecords were then filtered using an acausal bi-directional, eighth-order Butterworth filter. Fordigital records, low-cut filter frequencies weredetermined by considering the signal-to-noise ra-tio between each channel and a model of thenoise obtained from the pre-event memory. Sincethis type of model does not account for “signal-generated” noise (Boore and Bommer 2005), theresults were checked through visual inspection

of the velocity and displacement traces ob-tained from integration of the filtered acceler-ation record. Visual inspection of these traceswas also the basis for the selection of the low-cut filter frequency when no pre-event memoryof digital records was available. For analoguerecords, fixed traces were not available to allowthe identification of low-frequency noise. There-fore, the Fourier Amplitude Spectrum (FAS)of the unfiltered accelerogram was comparedwith the noise spectrum estimated from studiesof instruments and digitising apparatus such asthose proposed by Lee and Trifunac (1990) andSkarlatoudis et al. (2003), which were used asguide for the selection of low-cut filters. Sincethese studies correspond to a particular combi-nation of accelerograph and digitiser, which doesnot correspond to that of data being processed, vi-sual examination of the velocity and displacementtraces was also used as a basis for the selection ofthe low-cut filter frequency.

In selecting low-cut filter frequencies, the filterparameter was chosen to give a signal-to-noiseratio of 2. It is noted that the comparison of theFAS of the record with that of the noise indicatesthe ratio of signal-plus-noise to noise, hence if thedesired target is a signal-to-noise ratio of 2, theratio of the record FAS to that of the noise modelshould be 3. The maximum usable period of thespectrum was then defined as 0.8 times the low-cut filter period, as suggested by Abrahamson andSilva (1997), which is broadly consistent with thelimits suggested by Akkar and Bommer (2006).On this basis, it was decided that for about 85%of the records included in the database, the spec-tral ordinates could be reliably calculated up to3 s (or up to 4 s or longer for digital records),although for a few analogue accelerograms, theusable period range could only be extended up to2 s. Finally, removal of high-frequency noise wasachieved by using high frequency cut-off filters at25 Hz for records from analogue instruments and50 Hz for records from digital instruments. Peakvalues of acceleration and velocity and accelera-tion response spectra values for 5% of the criticaldamping were then obtained from the processeddata.

The other relevant parameter for each recordis the source-to-site distance. The source-to-site

38 J Seismol (2011) 15:19–41

distance was characterised in terms of the closestdistance to the earthquake fault plane or rup-ture distance (Rrup). Fault plane dimensions andorientations were obtained from published finite-source rupture models when available (e.g., Abe1972; Hartzell and Langer 1993; Mendoza et al.1994; Choy and Dewey 1988; Pritchard et al. 2007;Ji and Zeng 2007). Events for which fault-planegeometries from finite-fault inversion were avail-able have moment magnitudes 7.1 ≤ MW ≤ 8.4and contribute 70% of the strong-motion dataincluded in the database. For the 1966 (MW 8.1)and 1970 (MW 8.0) Peruvian earthquakes, therupture areas assumed for source-to-site distancecomputations were those estimated by Abe (1972)based on early aftershocks distributions; thesetwo events only contribute two records. For theaftershocks of the 1985 Valparaiso event, withmoment magnitudes 6.3 ≤ MW ≤ 7.1, the circu-lar rupture geometries determined by Choy andDewey (1988) were used to estimate the corre-sponding rupture distances.

For the remaining events, for which neitherfinite-source models nor reliable distributions ofearly aftershocks were available, an alternativeapproach was used to estimate the distance met-rics. Fault-rupture dimensions were estimatedfrom empirical relationships between rupture areaand moment magnitude (MW) for interface andintraslab-type events that have been determinedin Strasser et al. (2010). The rupture plane wasthen located in space, assuming that the epicentrelies above the centre of a dipping plane. Thestrike, dip and rake of the fault plane were as-sumed to correspond to the preferred focal planeof the two sets of angles listed in the HarvardCMT catalogue. On the other hand, for intraslab-type events the main focal plane was assumed tobe that suggested by individual studies of theseevents. For instance, for the 15 October 1997 (MW

7.1) Punitaqui event, the orientation of the actualfault plane was estimated to be the almost verticalnodal plane of the two sets of angles reportedin the Harvard CMT catalogue, based on thebody-wave modelling for this event carried out byLemoine et al. (2001). Similarly, the orientationof the preferred focal planes for the 7 November1981 (MS 6.7) and the 5 January 1974 (MS

6.7) events used herein were those suggested by

Astiz et al. (1988) and Langer and Spence (1995),respectively. This approach is expected to bea reasonable approximation for the purpose ofsource-to-site distance calculations in view of thefact that most of the events for which this as-sumption was applied correspond to intraslabevents with magnitude 5.9 ≤ MS ≤ 6.8, whichwere recorded at large distances and thus theirfault dimensions are not likely to be very largecompared to the source-to-site distances.

3 Conclusions

A database of strong-motion recordings for thePeru–Chile region from 1966 to 2007 has beencompiled, with particular emphasis on the qual-ity of both the data and the metadata associatedwith the recordings. The development of reliableregional strong-motion databases for subductionevents is of prime importance as they increasethe confidence in the results of both regionaland global ground-motion prediction studies forsubduction regimes. While the database presentedin this study is not sufficient for the derivationof a new predictive equation for ground motionsfrom subduction-type events in the Peru–Chileregion, it significantly expands the global databaseof strong-motion data and associated metadatathat can be used in the derivation of predictiveequations for subduction environments. Indeed,the compiled database further extends the magni-tude range of the currently available global data-bases for interface events (e.g., Youngs et al. 1997;Atkinson and Boore 2003, 2008) by the inclusionof data from the 2001 (MW 8.4) Peruvian eventand further supplements the global intraslab databy the inclusion of recordings from the 2005 (MW

7.8) Chilean event. Although the database pre-sented in this paper includes strong-motion datarecorded from 1966 to 2007, the present work canbe easily extended to include more recordings andmetadata from this region as they become avail-able, including those from the MW 8.8 earthquakethat struck Chile on 27 February 2010, as thisstudy was being finalised.

The compiled database will also allow the as-sessment of the existing predictive models forsubduction-type events in terms of their suitability

J Seismol (2011) 15:19–41 39

for this region, which directly influences seismichazard assessment in this region.

Acknowledgements The doctoral research of the firstauthor, on which this paper is based, has been partiallyfunded by the Alβan Programme of the European Unionunder scholarship E05D053967CO and the COLFUTUROprogramme; their financial support is gratefully acknowl-edged. The authors thank Bertrand Delouis for providinginformation on the source process of the 13 June 2005Chilean earthquake.

We are grateful to two anonymous reviewers fortheir constructive feedback that helped to improve themanuscript.

References

Abe K (1972) Mechanics and tectonic implications of the1966 and 1970 Peru earthquakes. Phys Earth PlanetInter 5:367–379

Abrahamson NA, Silva WJ (1997) Empirical responsespectral attenuation relations for shallow crustalearthquakes. Seismol Res Lett 68(1):94–127

Aguilar Bardales Z, Alva Hurtado J (2007) Seismicmicrozonation of Lima. In: Proceedings of theinternational conference on earthquake engineering,Centro Peruano Japonés de Investigaciones Sismicas yMitigación de Desastres (CISMID), UniversidadNacional de Ingeniería, Lima, Peru. Available onlineat: http://www.cismid.edu.uni.pe/descargas/redacis/redacis32_p.pdf. Last accessed November 2009 (inSpanish)

Akkar S, Bommer JJ (2006) Influence of long-period filtercut-off on elastic spectral displacements. Earthq EngStruct Dyn 35(9):1145–1165

Alva Hurtado J (2005) Sismo de Intraplaca ocurrido enTarapacá el 13 de Junio del 2005. Presentation at theCentro Peruano Japones de Investigaciones Sismicasy Mitigacion de Desastres (CISMID), UniversidadNacional de Ingenieria, Lima, Peru (in Spanish)

Araneda C, Saragoni GR (1994) Project of geological sur-vey of strong motion site in Central Chile. Report forthe Kajima Institute of Construction Technology ofTokyo, Santiago, Chile

Astiz L, Lay T, Kanamori H (1988) Large intermediate-depth earthquakes and the subduction process. PhysEarth Planet Inter 53(1–2):80–166

Atkinson GM, Boore DM (2003) Empirical ground-motion relations for subduction-zone earthquakes andtheir application to Cascadia and other regions. BullSeismol Soc Am 93(4):1703–1729

Atkinson GM, Boore DM (2008) Erratum to: Empiricalground-motion relations for subduction-zone earth-quakes and their application to Cascadia and otherregions. Bull Seismol Soc Am 98(4):2567–2569

Barazangi M, Isacks BL (1976) Spatial distribution ofearthquakes and subduction of the Nazca platebeneath South America. Geology 4(11):686–692

Beck S, Barrientos S, Kausel E, Reyes M (1998) Sourcecharacteristics of historic earthquakes along the cen-tral Chile subduction zone. J South Am Earth Sci11(2):115–129

Bernal I, Tavera H (2007a) Peak accelerations recorded inthe City of Lima. In: The August 15, 2007 (7.0 ML)Pisco Earthquake (Preliminary report). Direcciónde Sismología, Instituto Geofísico del Perú (IGP)(in Spanish)

Bernal I, Tavera H (2007b) Peak accelerations recordedin the City of Ica. In: The August 15, 2007 (7.0ML) Pisco Earthquake (Preliminary report). Direc-ción de Sismología, Instituto Geofísico del Perú (IGP)(in Spanish)

Boore DM (2008) TSPP: a collection of FORTRAN pro-grams for processing and manipulating time series.Available online from http://www.daveboore.com/software_online.htm. Last accessed December 2009

Boore DM, Bommer JJ (2005) Processing of strong-motionaccelerograms: needs, options and consequences. SoilDyn Earthqu Eng 25(2):93–115

Boroschek R, Comte D (2006) Amplitude and fre-quency characteristics of the 2001 Southern Peru,Mw = 8.4 earthquake records. J Seismol 10(3):353–369

Cahill T, Isacks BL (1992) Seismicity and shape of the sub-ducted Nazca plate. J Geophys Res 97(B12):17503–17529

Campbell KW, Algermissen ST, Kausel E, Highland LM(1989) Processed strong-motion data for the CentralChile earthquake of March 3, 1985: fifteen accelero-graph sites owned by CHILECTRA, ENDESA,and the Department of Geology and Geophysics,University of Chile. Open-File Report 89–448, US Ge-ological Survey, 328 pp

Campbell KW, Algermissen ST, Kausel E, Highland LM(1990) Processed strong-motion data for the CentralChile aftershock of April 9, 1985: nine accelerographsites owned by CHILECTRA, ENDESA, and theDepartment of Geology and Geophysics, Universityof Chile. Open-File Report 90–46, US Geological Sur-vey, 171 pp

Çelebi M (1987) Topographical and geological am-plifications determined from strong-motion and after-shock records of the 3 March 1985 Chile earthquake.Bull Seismol Soc Am 77(4):1147–1167

Çelebi M (1988) Processed Chile earthquake records of3 March 1985 and aftershocks. Open-File report 87–195, revised October 1988, US Geological Survey,254 pp

Choy GL, Dewey JW (1988) Rupture process of an ex-tended earthquake sequence: teleseismic analysis ofthe Chilean earthquake of March 3, 1985. J GeophysRes 93(B2):1103–1118

Cifuentes IL (1989) The 1960 Chilean earthquakes. JGeophys Res 94(B1):665–680

Comte D, Pardo M (1991) Reappraisal of great historicalearthquakes in the northern Chile and southern Peruseismic gaps. Nat Hazards 4(1):23–44

Comte D, Suárez G (1995) Stress distribution and geom-etry of the subducting Nazca plate in northern Chile

40 J Seismol (2011) 15:19–41

using teleseismically recorded earthquakes. GeophysJ Int 122(2):419–440

Comte D, Eisenberg A, Lorca E, Pardo M, Ponce I,Saragoni R, Singh SK, Suárez G (1986) The 1985Central Chile earthquake: a repeat of previous greatearthquakes in the region? Science 233(4762):449–453

Comte D, Pardo M, Dorbath L, Dorbath C, Haessler H,Rivera L, Cisternas A, Ponce L (1994) Determinationof seismogenic interplate contact zone and crustal seis-micity around Antofagasta, northern Chile using localdata. Geophys J Int 116(3):553–561

Cortez-Flores A (2004) Site response of the 2001 SouthernPeru earthquake. MSc Thesis, Washington StateUniversity

Delouis B, Legrand D (2007) Mw 7.8 Tarapaca interme-diate depth earthquake of 13 June 2005 (northernChile): fault plane identification and slip distributionby waveform inversion. Geophys Res Lett 34(1):Arti-cle No. L01304

Delouis B, Cisternas A, Dorbath L, Rivera L, Kausel E(1996) The Andean subduction zone between 22 and25◦S (northern Chile): precise geometry and state ofstress. Tectonophysics 259(1–3):81–100

DeMets C, Gordon RG, Argus DF, Stein S (1990) Currentplate motions. Geophys J Int 101(2):425–478

Dorbath L, Cisternas A, Dorbath C (1990) Assessment ofthe size of large and great historical earthquakes inPeru. Bull Seismol Soc Am 80(3):551–576

Douglas J (2003) What is a poor quality strong-motionrecord? Bull Earthquake Eng 1(1):141–156

EERI (1986) The Chile earthquake of March 3, 1985—EERI reconnaissance report. Earthq Spectra 2(2):249–508

EERI (2007) Learning from earthquakes: the Pisco, Peru,earthquake of August 15, 2007. EERI Special Earth-quake Report, EERI Newsletter, October 2007

Engdahl ER, Villaseñor A (2002) Global seismicity:1900–1999. In: Lee WHK, Kanamori H, JenningsPC, Kisslinger C (eds) International handbook ofearthquake and engineering seismology. Academic,New York, pp 665–690

Engdahl ER, van der Hilst R, Buland R (1998) Global tele-seismic earthquake relocation with improved traveltimes and procedures for depth determination. BullSeismol Soc Am 88(3):722–743

Espinosa AF, Husid R, Algermissen ST, De las Casas J(1977) The Lima earthquake of October 3, 1974; in-tensity distribution. Bull Seismol Soc Am 67(5):1429–1439

Hartzell S, Langer C (1993) Importance of model para-meterization in finite fault inversions: application tothe 1974 Mw 8.0 Peru earthquake. J Geophys Res98(B12):22123–22134

International Seismological Centre (2009) EHB Bulletin.http://www.isc.ac.uk/search/bulletin/ehb.html. Last ac-cessed May 2009

Ji C, Zeng Y (2007) Preliminary results of the August 15,2007 Mw 8.0 coast of central Peru earthquake. USGeological Survey web report. Available online athttp://earthquake.usgs.gov/eqcenter/eqinthenews/2007/

us2007gbcv/finite_fault.php. Last accessed November2009

Jordan TE, Isacks BL, Allmendinger RW, Brewer JA,Ramos VA, Ando CJ (1983) Andean tectonics relatedto geometry of subducted Nazca plate. Bull Geol SocAm 94(3):341–361

Kausel E (1986) Los Terremotos de Agosto de 1868 yMayo de 1877 que afectaron el Sur del Peru y Nortede Chile. Bol Acad Chil Cienc 3:8–12

Kelleher JA (1972) Rupture zones of large SouthAmerican earthquakes and some predictions. J Geo-phys Res 77(11):2087–2103

Langer CJ, Spence W (1995) The 1974 Peru earthquakeseries. Bull Seismol Soc Am 85(3):665–687

Lee VW, Trifunac MD (1990) Automatic digitization andprocessing of accelerograms using PC. Report CE-90–03, Department of Civil Engineering, Universityof Southern California, Los Angeles, California,115 pp

Lemoine A, Madariaga R, Campos J (2001) Evidencefor earthquake interaction in Central Chile: the July1997–September 1998 sequence. Geophys Res Lett28(14):2743–2746

Lemoine A, Madariaga R, Campos J (2002) Slab-pull andslab-push earthquakes in the Mexican, Chilean andPeruvian subduction zones. Phys Earth Planet Inter132(1–3):157–175

Lomnitz C (2004) Major earthquakes of Chile: a historicalsurvey, 1535–1960. Seismol Res Lett 75(3):368–378

Luppichini N (2004) Interpretación de los acelerogramasdel terremoto de Chile Central de 1985 considerandoondas sísmicas de alta frecuencia. Civil EngineeringDissertation, University of Chile

McVerry G, Zhao J, Abrahamson N, Somerville P (2006)New Zealand acceleration response spectrum attenu-ation relations for crustal and subduction zone earth-quakes. Bull N Z Soc Earthq Eng 39(1):1–58

Mendoza C, Hartzell S, Monfret T (1994) Wide-bandanalysis of the 3 March 1985 central Chile earth-quake: overall source process and rupture history. BullSeismol Soc Am 84(2):269–283

Midorikawa S (1992) Site effects on strong-motion recordsof the 1985 Chile earthquake and their non-linear be-haviour. In: Proceedings of the 10th World Conferencein Earthquake Engineering, vol 2, Madrid, Spain,19–24 July, pp 1031–1036

Midorikawa S, Riddell R, Cruz E (1991) Strong-motionaccelerograph array in Santiago, Chile, and prelimi-nary evaluation of site effects. Earthq Eng Struct Dyn20(5):403–407

NEHRP (1997) Recommended provisions for seismicregulations for new buildings and other structures.Report FEMA 303, US Federal Emergency Manage-ment Agency, Washington, DC

Nishenko SP (1985) Seismic potential for large and greatinterplate earthquakes along the Chilean and southernPeruvian margins of South America: a quantitativereappraisal. J Geophys Res 90(B5):3589–3616

Pacheco JF, Sykes LR (2002) Seismic moment catalog oflarge shallow earthquakes, 1900 to 1989. Bull SeismolSoc Am 82(3):1306–1349

J Seismol (2011) 15:19–41 41

Pardo M, Comte D, Monfret T, Boroschek R, Astroza M(2002a) The October 15, 1997 Punitaqui earthquake(Mw=7.1): a destructive event within the subductingNazca plate in central Chile. Tectonophysics 345(1–4):199–210

Pardo M, Comte D, Monfret T (2002b) Seismotectonicand stress distribution in the central Chile subductionzone. J South Am Earth Sci 15(1):11–22

Pritchard M, Norabuena E, Ji C, Boroschek R, ComteD, Simons M, Dixon T, Rosen P (2007) Geodetic,teleseismic, and strong motion constraints on slip fromrecent southern Peru subduction zone earthquakes.J Geophys Res 112(B3):1–24

Riddell R (1995) Inelastic design spectra accounting forsoil conditions. Earthq Eng Struct Dyn 24(11):1491–1510

Rodriguez-Marek A, Bray JD, Abrahamson N (2001) Anempirical geotechnical seismic site response proce-dure. Earthq Spectra 19(3):653–675

Rodriguez-Marek A, Hurtado J, Cox B, Meneses J,Moreno V, Olcese M, Sancio R, Wartman J (2007)Preliminary reconnaissance report on the geotechni-cal engineering aspects of the August 15, 2007 Pisco,Peru Earthquake. Report of the National ScienceFoundation-Sponsored Geotechnical EarthquakeEngineering Reconnaissance (GEER) Team. Web re-port available from: http://gees.usc.edu/GEER/Peru_2007/Peru_2007_WebPage/index.htm. Last accessedNovember 2009

Rosenblad BL, Bay JA (2008) Shear wave velocity pro-files determined from SASW measurements at sitesaffected by the August 15th, 2007 earthquake in Peru.Report prepared for CERESIS, January 2008, 60 pp

Ruiz S, Saragoni G (2005) Attenuation equations forsubduction-zone earthquakes in Chile considering twoseismogenic mechanisms and site effects. Proceedings,IX Jornadas Chilenas de Sismología e Ingeniería An-tisísmica, Concepción, Chile, 16–19 November 2005.Paper No. A01–15 (in Spanish)

Skarlatoudis A, Papazachos C, Margaris B (2003) Deter-mination of noise spectra from strong motion datarecorded in Greece. J Seismol 7(4):533–540

SNGM (1982) 1:250,000 geologic map of Chile. ServicioNacional de Geologia y Mineria, Chile

Stauder W (1975) Subduction of the Nazca plate underPeru as evidenced by focal mechanisms and by seis-micity. J Geophys Res 80(8):1053–1064

Strasser FO, Arango MC, Bommer JJ (2010) Scalingof source dimensions for interface and intraslabsubduction-zone earthquakes with moment magni-tude. Seismol Res Lett (in press)

Suárez G, Comte D (1993) Comment on ‘Seismic cou-pling along the Chilean subduction zone’ by BWTichelaar and LR Ruff. J Geophys Res 98(B9):15825–15828

Tavera H, Bernal I (2005) Spatial distribution of ruptureareas and seismic gaps in western Peru. Special volumeN6 commemorating Alberto Giesecke, GeologicalSociety of Peru, pp 89–92 (in Spanish)

Tavera H, Bernal I, Strasser FO, Arango-Gaviria MC,Alarcon JE, Bommer JJ (2008) Ground motions ob-served during the 15 August 2007 Pisco, Peru, earth-quake. Bull Earthq Eng 7(1):71–111

Tavera H, Buforn E, Bernal I, Antayhua Y, VilacapomaL (2002) The Arequipa (Peru) earthquake of June 23,2001. J Seismol 6(2):279–283

Tichelaar BW, Ruff LJ (1991) Seismic coupling alongthe Chilean subduction zone. J Geophys Res 96(B7):11997–12022

Wells DL, Coppersmith KJ (1994) New empirical relation-ships among magnitude, rupture length, rupture width,rupture area, and surface displacement. Bull SeismolSoc Am 84(4):974–1002

Youngs RR, Chiou SJ, Silva WJ, Humphrey JR (1997)Strong ground motion attenuation relationships forsubduction zone earthquakes. Seismol Res Lett 68(1):58–77

Zhao J, Irikura K, Zhang J, Fukushima Y, Somerville P,Asano A, Ohno Y, Oouchi T, Takahashi T, OgawaH (2006a) An empirical site-classification method forstrong-motion stations in Japan using H/V responsespectral ratio. Bull Seismol Soc Am 96(3):914–925

Zhao JX, Zhang J, Asano A, Ohno Y, Oouchi T,Takahashi T, Ogawa H, Irikura K, Thio HK,Somerville PG, Fukushima Y (2006b) Attenuation re-lations of strong ground motion in Japan using siteclassification based on predominant period. Bull Seis-mol Soc Am 96(3):898–913