The 1986 Kalamata (South Peloponnesus) Earthquake: Detailed Study of A Normal Fault, Evidences For East-West Extension in The Hellenic Arc

The 1986 Kalamata (South Peloponnesus) Earthquake:
Detailed Study of a Normal Fault, Evidences for

East-West Extension in the Hellenic Arc
H Lyon-Caen, R. Armijo, J Drakopoulos, J Baskoutass, N Delibassis, R.
Gaulon, V Kouskouna, J Latoussakis, K. Makropoulos, P Papadimitriou, et al.
To cite this version:

H Lyon-Caen, R. Armijo, J Drakopoulos, J Baskoutass, N Delibassis, et al.. The 1986 Kalamata (South
Peloponnesus) Earthquake: Detailed Study of a Normal Fault, Evidences for East-West Extension in
the Hellenic Arc. Journal of Geophysical Research, American Geophysical Union, 1988, 93, pp.967 -
982. �hal-01994159�
HAL Id: hal-01994159

https://hal.archives-ouvertes.fr/hal-01994159
Submitted on 25 Jan 2019
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 93, NO. B12, PAGES 14,967-15,000, DECEMBER 10, 1988
The 1986Kalemate(SouthPeloponnesus)
Earthquake'
Detailed Study of a Normal Fault,
Evidences for East-West Extension in the Hellenic Arc
H. LYON-CAEN,1 R. ARMIJO,2 J. DRAKOPOULOS,3'4 J. BASKOUTASS, 4

N. DELIBASSIS,
3 R. GAULON,1 V. KOUSKOUNA,3 J. LATOUSSAKIS, 4
K. MAKROPOULOS,3 P. PAPADIMITRIOU, 1 D. PAPANASTASSIOU,4 AND G. PEDOTTI5
Tectonic and seismological data collected in the field following the September 13, 1986,
Kalemateearthquake(southPeloponnesus)
are presentedand analyzedto discuss
the earthquake
rupture process and the regional tectonics. The event occurred on the Kalamata normal fault
whose trace was mapped with SPOT images and topographic and field observations. This fault
is part of an approximately NNW-SSE en •chelon system cutting through the Hellenic nappes.
The fault strikingN15øE on the average,with a dip of about50ø, has a minimumcumulated
Quaternary throw of the order of 1 km. The measured coseismic slip is 6-18 cm over a length of 6
km. Themainshock focalmechanism obtained fromlong-period
waveformmodeling (strike=201
ø
(+ 10o,_20o), dip=45oq_5o, rake=283o (+ 10o,_25o)) represents
almostpureeast-westextension
and is in good agreement with tectonic observations. The centroid depth is constrained to 5-t-3
kmandtheseismic
moment
to ?.0q_2.$x
1017N m. Over700aftershocks,
located
bya 16-station
network installed after the earthquake for a period of 2 weeks, define two clusters separated by
a "gap" of aftershock activity, from the surface to a depth of about 10 km. The main cluster,
to the south,definesa 45ø westdippingplanewhichlieson the downwardextensionof the fault
mapped at the surface. Focal mechanisms of aftershocks on this fault plane are homogeneous
and represent E-W extension as the main shock. In contrast, the majority of focal mechanisms in
the uppermost part of the footwall show more or less E-W compression, probably corresponding
to postseismic stress release. The northern cluster of aftershocks is very dense and located away
from the surface rupture, within a relay zone between the Kalemate and the next en •chelon
faults to the NW, the Thouria faults. There focal mechanisms represent extension from about
Nl15ø to N?0ø andN20ø, corresponding
mostlyto fault reactivationin an areawherenonrigid
deformations prevail. The main shock probably initiated in this relay zone 3-4 s before the
rupture front reached the main fault plane and released most of the energy there, the rupture
presumably propagating southward. The focal mechanism of the Kalemate earthquake and
that of the April 27, 1965, earthquake located to the northwest of Crete, as well as the regional
active normal fault pattern, imply that E-W extension oblique to the Hellenic arc is presently the
dominant tectonic regime. E-W stretching occurs partly on reactivated NW-SE faults parallel
to the Hellenic structures but mostly on newly formed N-S normal faults across those structures.
The latter faults are responsible for the apparent segmentation of the Hellenic belt from southorn
Peloponnesus to Crete. The existence of active E-W extension in this region implies a recent
change in the tectonic regime and consequently a change in boundary conditions at the subduction
r.one, probably in response to the incoming margin of Africa.
INTRODUCTION Because of the low level of seismicity compared with that

of areas nearby like the gulf of Corinth or Thessaloniki and
TheSeptember 13,1986,earthquake (ML=5.5(Athens), the apparently long recurrence time of medium- or large-
M8=5.8(National Earthquake InformationService,NEIS)) size earthquakes, the tectonic activity in this area has been
that destroyed a largepart of the city of Kalamatain somewhat overlooked up to now. In particular, no focal
southwestern Peloponnesus (maximum intensityX) is the mechanisms were available for this region. The detailed
first shallow event in the overridingcontinentalcrust of study of the Kalamata earthquake presentedhere bears both
theAegean between CorinthandCreteto bewellrecorded on the seismotectonics of the Aegean and on earthquake
teleseismically
andto haveproduced
clearsurface
faulting. rupture processes.
The Aegean area has long been recognized as a region of
intense and widespread continental extension, but the me-
1 Laboratoire
de Sismologie,
Institutde Physique
du Globe,
Paris, France. chanismsof this extensionare still a matter of debate[e.g.,
:2Laboratoire
deTectonique,
InstitutdePhysique
du Globe, McKenzie, 1978; Le Piehon and Angellet, 1979; Mercier et
Paris, France. al., 1979a;JacksonandMcKenzie,1983]. The directionof
3 Divisionof Geophysics-Geothermy,
University
of Athens, active extension is approximately north-south in Thessalo-
Greece.
4 Seismological of Athens, niki [SoufierisandStewart,1981;Mercieret al., 1979b;Hatz-
Institut,NationalObservatory
Greece. reid et al., 1987],in the gulf of Corinth [Jacksonet al., 1982;
5 IRIGM, Observatoire
deGrenoble,
France Sgbrier,1977] and probablyalsoin the Cyclades,the nor-
thern Aegean,and westernAnatolia lAngeliefet al., .1982].
Copyright 1988 by the American Geophysical Union. Such north-south extension, despite small episodic or local
Paper number 88JB03484. fluctuations, appears to be the dominant tectonic regime
0148-022?/ 88/ 88JB-03484$05.00 sincethe Pliocene [Mercier et al., 1979a; Angellet et al.,
14,967
14,968 LYON- CAEN ET AL.: THE 1986, KALAMATA, EARTHQUAKE
1981.2.25
CORiN1981.2.24
1981.3.4 •o
O o
A E G E A N
lam
S E A
OF x
MESSENIA
• LACONIA
ß ww
36"
•) ß
ß' (9
(9'(9
• [ 1'•., • 1965.4.27
ß
ß '(9
ß
ON
22" 28" 100kin
I
35
I
22 ø 23" 24"
LYON- CAEN ET AL.: THE 1986, KALAMATA, EARTHQUAKE 14,969
1982;Lyberis,1984]. In contrastwith this apparentunifor- late Helleniccompression

(lateMiocene-Pliocene?)
andsub-
mity in most of the Aegean and westernAnatolia, the active sequentPliocene-Quaternaryextension. Compressionmay
and Quaternary tectonics just to the north of the Hellenic have initiated the formation of internal relief by large-scale,
trench, from southern Peloponnesusto Crete and Rhodes,gentle NW-SE folding of the nappe structure. This led
to updoming of ranges at anticlinoria and downwarpingof
is more complex: both north-south and east-west extension
has beenobservedlAngeliefet al., 1982]. However,local basinsat synclinoria[Institute of Geologyand Mineral Ex-
microtectonic
measurements lAngelief,1979]and Landsat ploration(IGME), 1986]. Extensionapparentlyamplified
imageinterpretation[Armijoet al., 1986](Figure1)suggest this smooth, "compressional" relief by large, often en •che-
that the most recent movements are being accommodated lon normal faults oriented approximately N-S which roughly
by north-south normal faulting and east-west extension. broke at the boundaries between ranges and basins and pro-
In this paper we present both tectonic and seismologi- duced more localized uplift and downdrop on them. Parti-
cal data collected in the field following the September 13, cularly large Quaternary and active normal faults are clearly
1986, earthquakein Kalamata. We mappedthe Kalamata visible on satellite images along the two edges of the Tay-
fault and other active faults in the area with the aid of air getosrange(Figure 1) JArmOoet al., 1986]but havenot
photographs,topographicmaps, and high-resolutionSPOT been described in detail so far. The September 13, 1986,
images. We evaluatedboth the long-term(structuraland earthquake broke along one of those normal faults whose
morphological)offsetsand the 1986 coseismicslip. We trace runs along the western mountain front of the Tayge-
constrain the focal mechanism and the centroid depth of tos massif,only about 4 km east of Kalamata (Kalamata
the main shock using waveform modeling and present an fault, Plate I and Figures2, 3, and 4) (Plate I is shown
analysisof the locationsand individualfocalmechanisms
of here in black and white. The color version can be found in
aftershocks recorded by a portable network of 16 stations the separatecolorsectionin this issue). No surfacedisrup-
for a periodof 10 days. Papazachos et al. [1988]already tion or earthquake had been observed and associatedto this
described the main shock focal mechanism from first motion Quaternary fault before the 1986 event.
readingsand the resultsof a 3-day aftershocksstudy with
a network of five portable stations. The quality and large SurfaceFaulting of the September13, 1986, Earthquake
amount of data enable us to constrain precisely the geome- Small but sharp surface faulting occurred along the trace
try of the aftershockzone. We then combineseismological of the Kalamata fault during the earthquake. This was
and tectonic observations for a discussionof both the rup- immediately noticed by the inhabitants because the main
ture process of this earthquake and the regional east-west
fault break crossedthe road to Eleochori(site 5 in Figure
extensional tectonics.
2b). We wereableto followthe breaksalmostcontinuously
TECTONIC OBSERVATIONS in the field for about 4.5 km between sites I and 6. The
breaks, however, extend with some discontinuitiesup to site
Quaternary Tectonic Setting 8 (Figure2b), reachinga total lengthof 6 kin. They usually
The structure and topography of southern Peloponnesus produce a fissure and a normal fault step in colluvium or
is markedby a seriesof threehigh (1200-2400m), long(50- terraced olive orchards which lie against bedrock outcrops
100kin), approximately NNW-SSEranges,separatedby the on the mountainfront (Figure 5a). The coseismic fissure
and stepare generallyvery steep(70ø-90ø),about6-18 cm
two parallel depressionsof the Pamisos and Eurotas rivers
(Figure 1). Thoserangesand basinsextendsouthward high and 2-6 cm wide on the average(Figure5). They are
into the Aegean Sea by pronounced peninsulas separated typically located a few to several tens of centimeters west of
by the gulfs of Messenid and Laconia. The bulk of this large normalfault slickensidesin the bedrock(Figure 5d).
internalrelief (_• 3500 m) postdatesthe MioceneHellenic The step mergeswith those slickensideslocally along strike
nappe eraplacement, apparently as a consequenceof both (Figures
5band5d)andalsocertainly
at depth(_• 1 m).
Thus the seismic break branches off the master fault near
the surface where the reduction of soil cohesion facilitates
Fig. 1. (opposite)Seismotectonic map of the Pelopon- the rupturepropagationalongsteeperplanes(Figure6).
nesusand adjacentareas.Topographyand bathymetry(in The normal fault slickensideslie at the main slopebreak at
meters) are from the InternationalBathymetry Chart of the base of the front of Mount Kalathio which is the uplifted
the Mediterranean[UNESCO,1981]at 1:1, 000, 000 scale. footwall visible in Figure 4. Most of the mountain front is
Contours are at 600 and 1200 m. Active faults are modified composed of crystalline Mesozoic-Eocene limestones of the
fromJackson et al. [1982]andLyberiset al. [19821
andfrom Ionianplatform(plattenkalk)whichhasbeenexhumedfrom
Landsatimageinterpretation[Armijoet mi.,1986].Inferred below the Tripolis nappe by normal faulting with a total
faults are shown segmented. All faults are mainly normal throw of the order of 1-2 km (Figures2a and 3a). With
and the tics point to the downdropped compartment. Fault regard to the geologyand the morphologythe total throw on
planesolutions of shalloweventsare from McKenzie[1972] the Kalamata fault reaches its maximum at Mount Kalathio.
for the April 5, 1965event,from Jacksonet al. [1982]for the North of Mount Kalathio the total throw diminishes rather
1981events,from this studyfor the April 27, 1965 (Figure abruptly and vanishessouthof Rents Rachi (Figures2 and
A1) and the September13, 1986,(Figure9a) events.Inset 4). The large dimensionof steep,youngtriangularfacets
showsthe tectonic framework and earthquakes with magni- onMountKalathiofront(_• 30ø, up to 250m high)andthe
tude larger than 5.0 in the Aegean region from 1900-1986. observationthat they derive from still larger facets which
Circles are for events less than 40 km deep; squaresare for form the overall,trapezoid-shapedmountainfront (700 m
events more than 40 km deep. Earthquakes are classifiedin high on the average,Figures4 and 2b) imply that muchof
four groupsof magnitude(M•5.5, 5.5•M•6.5, 6.5•M•7.5, the total throw acrossthe Kalamata fault is of Quaternary
7.5<M<8.5) corresponding to the differentsizesof symbols. age.
14,970
LYON-CAENET AL.:THE 1986,KALAMATA,
EARTHQUAKE
::•:•*!•x•:.•::•
.... J:.½:-
. :. "• :-'.•-----.•..,;....•
.-•-•::
..:.::
,.. ":• •::::•:•:::
.............
:::::::::::::::::::::
..:•::•:•:::-::•*-•-::'•-•::•
.•: ..•.:,. ::•'[.•
:':;•.:**: :'"'k.:..•r"'....•
........................
• •,'"•
.• ...
"•! ........
:........
•::•.'.
•;,,.-'-'•'------..: .....?/--.-:[...--
•.:
......
::.;.:::•--.:.--[--.:./:•
::.;•
...........
..,,,-,
.......
•.':.•...•
:- .•..$;•%:;;:::::.::•.
..:'.-•?':-:•'
........{- -..•.
,We•::
4- .•-.,>.-
...-..:.
..![:•::
.........
[' ...•.
:-*.....-:...,.
..??-..:
.........
......
•,
. ...:=...-..
....•
:.'-...::............
,?.,:..,-s:-:-•
½:-..:;..- :..'. ';,,,;"
..-•...:::....:::..".
.. .. *:':,:':..;:3.;..'.;;'
. ..:;•,•.!'. . . .-:r:;....
:....
ß.'L.:::
........... :.:½::*** ½;•:;½':::::?'½ . .;..*
!•:::;F:-•'..•::..
:.>::'f:.;--
• :•:::
½ ::•;•::•.; .....
•:,•..r..
6 *½
...............
':::::*:::'"*
.........
.',..,.-.-.½"t•;:'•.
;[;½
......
'•t•'*:"*:":;i.,.•
'"'•...........
'%. ,J:
•"*:
..'/f:;.-"'
......................
½-:'•*:'.
;• ./[• •.... "?; '.:::-:.•, :":'""• •.:-.,' '::• 27""2'....•...•........:.::...-;;
;::½ ....... - :' ' '.5 '
?:-':::'*.*•
:'•
ß
•:•. ..... :"'
:.....ii.'-'A
.--%
,..'/' .... •':;:-
.-•-';:':'"".•
4.:;
'....:,.....•
--;'
...;•:-•;•.'
?"'?;
'*?.,-,:"•" "':':"'-•'::--'
..... .;-':[::/.'•" a?'?•
• •.??'.':'""•;
......... :'::.":i:!:.!
-.:.•{.Z::.i;
' . ............
6-:"""
.,: ::-:"/'/•:-'/"'"•
.........
i"'S.::/'
-•"'
::i'..:.....
' ..../.
' ::•.:
-,, ;............
.x:So'
...... '."*•.
:.".'..'
.::,::'.-½
.:....":
77.'.::.:'"":::/;.';:
"c.'.... .
.'[[:'(--.-,::/'•:•:'".
4.':::'.......- .....
:'*:''[..•:i• '-'•' '• ?.•ry"-•.•....

....
.' ...........
....'" '....:.•-.:.
.• ....... :'•"•*:;...;•-"".:.•-:. ,.•*:-,•:
:;/?::•:.•-.-:• ......
:;.,,...•%.•' •"' • .:,•'".-'•:.'.:" '.":::'::.•
.......................................
::.....::•}'•;; -..............
' •. -::•..•*
•:..... '•?.•..............................
:i• .............. ......
---•-
........ .:•--:-:----½t-
--•:--.
•---•:s4:--::
.d•,-y......
*-. •...... :t•: •..:•-::::-'•--•-.
... •:-,:----:•::-
-<--,.......:,a
....
:• :, *****::-.
-•.,.,• '.•..'• ,.a .....
,-• •::: *......•*":•:'--•T'"';'•"•
•:--'-rla:-..*::'-"
•..•':•:-::'i:.:::.
.....;*!!!:.::?:(
......:."%:':,•
.......
..::.•;•':'*:-'-•:::f:½•!-:.!.•::.,?!f'"*"
.'. ..-.::..:-:::a•'::•,
......
•'- :•:: :::.• ...,,½
..... ---,•:.: ::•:::
.:--•. •:'"•-"*'-•-------
-••;"
....
;;........ ? ?*
......
i .............. '.... -.•
...... ,•* •.•:-,..•:::::.;,..:.•
..... ........
;'ca;: .... ;•"';•'::a
•'.; ............... '":'•*"*'•
'*': •* '"'
*':'*:•
.........
i :%
.... ':•;•'•'*":
.... '%::.."
i• r•, ..:-'"':'*'::'-•'?.•;•;::
':•!i•! ......
.........'"'-*':a•;'
'" %:'•. •,....:•:•!,
"•/.5:.*•...........
*":'•'•;•:•/!!.:*•:•J•½.:'::;•,';
- - '*-:':•-•-•"• ,•.., :-:::;:'::*•:'t•::::
.........
"'"•'"'"'"'"•'•?.•.•
.%...•-t::• :i'""'*'"• '•'•.
*:':
.........
'•'"'•'"•:'••••••••••;•:•
• •'*'•::•
'"" .:.,'........... i'::•:.'
•.
•
**-"%.•: :':"':"*::•"
..•.. • .... ::"-:
' :•::•i':"* -'.'""•'
*'"• ;;;•::::•::,-.•'•-•::'.•:•'• ::::-".•<;--'%:'""•:-•
ßd[;•:•:.,:
............
.................
•x• .';•::'
'*,,•':?'"
'-••:•:*
.:• ..... , :.•:.:, .-•_•...:•..:.:.,•½.•
,.•.:::,:
................................ ..::.:
..... .,...• ,:,..:..•
..• ......,• •..
• (• ........ ...:**,:!•:•
*........ .. .....
....."-
.;'• • ...•.................................
..................... .::..::.:.,::
......................................
:• .,.:
:: .. .•...**
':-%...: ...
•:.,::.**i•
•.... ......:.-.,•,:-----:
................
'.::-a.•;;•:::•½,
:'•': ......; ....... ..ß ;........
.
• ::.-:•':':%.*
............. 'T• ,**,-*',*:.,.-'-'..:::•'*
..... .:":•-•::•.:..:: !%..-•.•;•.<.•::•-';
.-:...:..•?*::'•',:a:•
.....'•'
.....
'"'::'":•:•
::• ............ ......

•'-•'•.. :!!:';;;:½;•:"½;F"-:'::-'"
';'"•:'•"
'•:• :::•U f'•'•:.•....... '•'"'•*"•*'•
...... .::.•::..;:.::•i•;$t•
.: .;;..'•.•:
,• ....•f':• ..•'"'; ' ?•:"------:-:-::

;•!:!:::-•-*::'*•:':':'•*•ii::i:
;• . :':•':::':".'.?'"
:;;2;:;•11:::•:':':•i•;;;:.;7';'".
........
;..
....
•:..'•....:;!:• '...................
•:z"?:::'•W'"'•/:•;S.......d.i:.;•_...._..?•.:
'•$.3::'"'
""'"".. •: ::•
.:•-•-•:*-'::'•:':':•:•!:*?*':T'"•:'•::;•'--...•;•;.'.•'..-.•*..•;i•..
..;,-"'
....... d.<.
'- •.,:..,,
;•::$::•
•..... •,:x**-.:.-.,. . •::•;,.-
,.,f ......
..'"'•':•'•!':;::.•?s,.c':•.'4:/
.........
•.........• .............•.•**-••••:••-..•_••••::.
:,-...•,::
:'.-:'
"•*f•
........
•:'•';:;:i:::•;.'"'•-':"".....'•;•'"'•:-•
'••i;"-:•!•:•'*::4"c•
:...a:.:x•:
*::'*".
...........
............
•'
'.......
*&:.;½i';•
!;:.e..:..•':::•
...... '""*•"•'•j;:-•:•½:;..
.:'; ":':'"-'"".':•
" ............. ":'/• " ....
•:•;{%•x.:.;::-..i'iii•'"
.:':i;i
-$...-*'
..........
5 k.m "i '''.................................

! 22.":'i:
O'
Plate1. SPOTimage oftheKalamataarea.XSchannel
2 (0.61-0.068/•m).
(Thecolorversion
anda
complete
description
ofthisfigure
canbefoundintheseparate
colorsection
in thisissue.)
The coseismicmovement correspondssystematicallyto

of Messenia (Plate1 andFigure2). In particular,the
normalfaultingwith downdropo• the westernblock,al-
roadto Vergawasnot cut by faultreactivationduringthe
thoughsoildeformation maylocallyyieldhangingwall col-
earthquake.Similarly, northof site8 (Figure2b),there
lapseagainstfootwallsalientsandan apparentreversefaul-
is no evidenceof coseismic
surfacefaulting. In contrast
tinggeometry (Figure5½).Thenetcoseismic slipisdirectly
w•ththe southern endof the ruptureat site1, however,
measured onthemasterfaultplaneat placeswherethesteep
no Quaternaryfault traceextendsnortheastward beyond
stepin thesoilmerges
withthatplane(Figures
5band5d). site8 andacrossthe WNW-ESElimestone rangeof Renta
Althoughsurfacefaulting was very clear at site 6 west of Rachi, which thereforemarks the northernmostend of both
Eleochori (Figures 2b,7aand7b)anda smallstepwasstill the seismicruptureand the quaternarybasinof Kalamata
visibleat site 8 (Figures2b and 7½),no localizedsurface (Figures
2band7a). Thenextmajornormal faultsystem
disruption was found between these two sites. There the bounding the Pamisos plain is found to be offset about
fault traceis markedby footwallfacetedspurscut in the
25 km to the WNW from the Kalamatafault trace,on
uppermost
Pindosnappe(Figures2a and 2b). The wa- the westernborderof the nearlyunfaultedGoupataRachi
ter dischargein springslocated on that trace at Perivolakia
plateau
(Thouria
fault,Plate1 andFigures
2aand3b).
(site7, Figure2b}isreported
bytheinhabitants
to havehad TheaveragestrikeoftheKalamata master
faultisN20øE,
large fluctuationsafter the earthquake. We concludethat andit isformedoftwofaultsubsegments:a N10øEstriking
betweensites6 and 8, the rupture was more diffusethan segment southof the bendat site3 anda N30øEstriking
southof site 6 or that it nearlyfailed to reachthe surface.
segment northof that bend(Figure2b). The dip measured
Southof site 1 (Figure2b) therewasno surfacebreak on metric slickensidesalong the master fault trace is 60ø-
althoughen 6chelon
Quaternary normalfaultingappears, 70øW(seeFigure5d). Thiscontrastswiththe shallower
on the SPOTimage,to extendsoutheastward
intothegulf dipsof about50øWobtainedby detailedmappingof the
.•Thouri
MESSENIA
KALAMATA
GULF OF MESSEN
7000 '
• Quaternary
2b
Pliocene
':"'"'":•
Pindos
unit
•'•
.....
"'.
-'..
'."*'*'-
ß Tertiary
.... Mesozoic Tripolis
'"'"':'"• unit
• Basement
Ionian platform
0 5 km
[ I I I I I I ?Kardhamili:
Fig. 2a. Map of activefaultsnearKalamata.Topography(in meters)and geologyare simplifiedfrom

the IGME, Athens, 1:50, 000 scale maps. Faults are compiled from the SPOT image in Plate 1, aerial
photographs, and field observations.Major faults are in bold. Nappe contacts are shown with triangles
pointing toward the upthrust unit. The box corresponds to Figure 2b. Sections a and b are shown on
Figures 3 a and 3 b.
fault trace acrossthe roughtopography(_• 200 m relief in their formationby continuinguplift of the footwall(Figure
canyons)(Figure 2b and 6). Coseismicslip at the surface 6).
was observed to be mainly normal with very small amounts Coseismic movements of 1986 on the Kalamata fault imply
of either right- or left-lateral component of slip. Such that it ruptured completely up to its northern extremity
undecided lateral components probably have little tectonic (site8) but that the breakdid not propagatealongthe whole
significance and may only reflect local influence of steep lengthof the fault system,southwardof site 1 (Figure2b).
slopes in transverse canyons, as in the vicinity of the road Except for local, small breaks on the road to Eleochori about
to Eleochori(site 5, Figure 2b). By contrast,the lateral 300 m west of site 5 and unclear but probable reactivation
component of slip in the slickensidesobservedat the base of of faults in the small graben south of Eleochori, none of the
the mountainfront is small(rake _>75ø) and consistentlysecondary faults mapped in Figure 2b was reactivated at
dextral (Figure 5d). Thoseslickensides are probablyof the surface. All other surface disruptions found in the area
Quaternary age, have formedwithin the bedrock/bedrock [Papazachos et ed.,1988]appearedto us to be gravitational
master fault contact at depth, and have been exhumed since slides with doubtful relation to active faults.
22o10 '
q001
%
KALAIVlATA
GULF
OF
MESSENIA
37o00 •' 37o00 '
0 1 2 km
I I I
22ø10 '
Fig. 2b. Detailed map of surface breaks of the September 13, 1986, earthquake. Circled numbers are
observation
spotson the main rupture(heavyline). Largemorphological
scarps(• 100 m high) and
triangular facets are hatched, Mount Kalathio front is in light grey. Surface breaks were not continuous
between sites 7 and 8. Other faults did not break during the earthquake.
The Kalamata fault is part of a larger fault system formed North of the Kalamata fault the offset with the Thouria fault
by en &chelon, west dipping normal faults at the border coincides with a Nl10øE deflection of the late Hellenic folds
betweenthe Taygetosrangeand the Pamisosplain (Figures in Pindosunits, whichusuallystrike N150øE, parallelto the
1 and 2a). South of the Kalamata fault the system generaltrendof the Taygetosrange(Figure2•). Thisfault
is complicated by an oblique, NW-SE trending graben at offset and fold deflection are probably located along a zone
Kampos and extend• probably into the gulf of Messenia by of lateral ramp in the Ionian unit and the overlying Tripolis
a largeoffshorefault westof Kardhamili(Figures1 and 2a). nappe (Figures2• and 3). It is thusvery likely that the
LYON- CAEN ET AL.: THE 1986,KALAMATA, EARTHQUAKE 14,973
l= o o o o o
I I
I I I I I I I
I I
I I I I I
I I I
I '
I
offset between the Kalamata and Thouria faults is located

along an older zone of transversefractures which do not
exhibit traces of surface reactivation during the Quaternary.
MAIN SHOCK
Location
Locationsof the main shock(September13, 1986,at 1724

GMT) givenby the EuropeanSeismological Center (CSEM)
andthe NationalObservatory of Athens(NOA) altnostcoin-
cide and are situated at the northeastern extremity of the
aftershock
zonefoundin this study(Table 1 andFigure8b).
The largestaftershock(ML--4.8, Athens)occurredon Sep-
tember15 at 1141GMT (followedby an ML--3.6 aftershock
on September15 at 1247GMT). Both aftershocks are loca-
ted to the south of the main shock, nearer to Kalamata city,
by CSEMandNOA (Table1). Forthe mainshockandthe
largestaftershock,
a singlethree-component
strongmotion
instrument(SMA-1 type) from the town of Kalamata was
in operation, and the record shows a time difference of 1.9 s
between the S arrival and the triggering time for the main
shock, and of 1.4 s for the aftershock. These data imply
that the main shock and the aftershock epicenters cannot
be much closer to Kalamata city than about 10 and 5.5 km,
respectively,
assuming a meanP velocitylargerthan 5 km/s,
a Vp/Vs ratio of 1.73and depthsof eventsof 5 km or more.
Qualitatively, these results are in agreement with locations
givenby NOA, CSEM andPapazachos et al. [1988],although
they differ from eachother by few kilometers(Table I and
Figure8b). Thesecalculateddistances wouldbe about25%
smaller(7.5 and 4.5 kin, respectively)if one assumes a 2-
km-thick sedimentary layer, with P and S velocities of 3.0
and 1.66km/s, beneaththe plainof Kalamata. We attemp-
ted to relocate the same two events with respect to locally
well-located aftershocks, but the data did not provide stable
relocations. One reason for this lack of resolution, at least
for the main shock, may be the apparent complexity of the
initial rupture. On teleseismic World-Wide Standard Seis-
mographNetwork(WWSSN) short-periodrecordsand on
short-period records at Global Digital SeismographNetwork
(GDSN) stations,this complexityis expressedas a smallar-
rival (equivalentto a magnitude4.5 or 5 event) about 3.5
s before the main event, suggesting that the rupture pro-
pagation was not smooth. The combinedinterpretation of
these short-period records and strong motion will be impor-
tant to better constrain the initiation and propagation of
the rupture on the fault (M. Besnardet al., manuscriptin
preparation,1989). Basedon the data discussed above,we
think that the relative location of the main shock and of
its largest aftershock with respect to the aftershock zone is
not very well constrained, but the main shock most likely
initiated within the northern half of the aftershock zone and
the September 15 aftershock within the central or southern
part.
It is interesting to note also that the centroid calculated
by Dziewon•kiet al. [1987](36.8øN,22.64øE)isshiftedal-
most 50 km to the east. Mislocations of about 50 km to the
northeast also occurred for the 1981 Corinth earthquakes
[DziewonMd
and Woodhouse,
1983]andfor the 1983Kephal-
linia earthquake. In contrast, centroid locations of earth-
quakes in the western Mediteranean like E1 Asham, 1980
(Algeria)[O•yedet al.,1983]or Constantine,
1985(Algeria)
[Bouni/et al., 1987]are preciseto about10 kin. It is likely
that these mislocationsreflect lateral heterogeneitiesin the
...
•X'- •,'L.-'"
'.
.. ,.,............
... • :*.- ., ..;/->.'"'.•,:':<,
..........:
<%'.7. ..........
'. };X.*.:'...,;.•}
.... :'....":.L....
';
..
Fig. 5. Detailsof thefaultbreaksouthof Eleochori

road(site5 onFigure2b).Figures5a, 5band
5c are at site 4; Figure5d is at site 2 (seeFigure2b). (a) Continuous,
3-cm-widefissuremakes10- to
12-cm-high stepin soil,lessthan I rn awayfromthe N45øEstrikingand60øWdippingfault planein
bedrock
(seeFigure6). (b) As stepin soilmerges
withbedrock/soil
faultplane,the net resolved
dip
component
of slip is 10 cm. (c) A 13-cm-high
step-and-fissure
3 rn awayfrom the bedrockfault. Because
of thecollapse
of the hanging
wall,the apparent
slipmaybe reverse.(d) Slickensides
andgrooves rake
82øNona N15øEstriking and70øWdipping plane.Only6- to 7-cm-widebarestripeat thebaseofthe
planewasexhumedby the downdrop of grass-covered
sediments
duringthe earthquake.Note that the
smallestbreakis observed
just at the footof MountKalathiowherethe long-termdisplacementreaches
a maximum.
WNW ESE
triangular facet
,-,., lOOm
I t 5O
piedmont sediments Cretaceous-Eocene
'plattenkalk'
(Ionian unit)
,-,., lm
slickensides
Cretaceous limestones
(Tripolis unit)
slip vector
bedrock
bedrock
Fig. 6. Schematic crosssectionof the Kalamatafault nearsite 2 (Figure2b). Circleblowsup the

surfacebreak geometry.Master fault dips about 50ø near the surface,slickensides
about dip 60ø, and
fissuresin soilwith lowcohesionarestill steeper(_• 80ø).
velocitystructurenot taken into accountby the preliminary pendentof the assumed

meanvelocityand of the depthand
reference
Earth model(PREM) [Dziewonski andAnderson, source duration.
1981]usedin centroiddeterminations.Insteadthey may By fixing oneof the focal mechanismparameterand inver-
result from either a high-velocity zone associatedto the sub- ting for the remainingones,we exploredthe uncertaintieson
ductingplate beneaththe Aegeanor a low-velocityzonere- eachparameter. The dip wasalwaysvery stable. The strike
lated to the widespreadextensionand high heat flow in the canvarybetween
185ø and220ø (seerows2-3onFigure9b)
Aegean or a combination of both. and the rake between290ø and 260ø withoutsignificantly
deteriorating
the fit of waveforms
(rows4-5 on Figure9b).
Waveform Analysis The centroiddepthis moderatlywell resolved,as expected
In order to constrain the depth and focal mechanism of for such a small and shallowevent, becauseof the trade-off
the main shock, we inverted available P and $H waveforms betweensourcetimefunctiondurationanddepth. However,
of long-periodWWSSN seismograms [Ndb•lek,1984; Me- SH waveformsallow constrainingthe depth to be lessthan
Caffre• andN•ibllek,1987]for focal mechanism parameters 8 kin. At largerdepth, the sSphasestarts to appearas a
(strike, dip, and rake), centroiddepth, seismicmoment, separatedarrival(at stationNAI andKBS in particular,see
and source time function. The inversion algorithm mini- row 6 on Figure9b). A very goodoverallfit of waveforms
mizes the difference between observed and calculated seis- and amplitudesis obtainedfor a depth of 5 km and a source
mogramms in a least squares sense. However, it is very durationof 4 s (Figure9a), yieldinga seismicmomentof
important to explore the range of acceptablevaluesfor focal 7.9x10
z7N m. Thusfora fixeddepththeseismic
moment
mechanism parameters and depth because the standard er- is well constrained. However, good fits can also be obtai-
rors provided by the algorithm do not representthe true un- ned for a depth of 3 km with a 5 s-longtime functionand a
certaintieson theseparameters[e.g. McCaffre•tandNdb•lek, seismicmoment 9.1x10z7N m orfora depthof 7 kmwith
1987]. Setsof calculationswere performedin a half-space a 2 s-longtime functionand a seismicmomentof 4.5x10z7
with a mean P wavevelocityvaryingfrom 5.4 to 6.0 kin/s, N m. We concludethat the depth is constrainedto 5+3kin
which correspondsto the acceptablerange deducedfrom the and the seismicmoment to 7.0+2.5 x 1017 N m.
analysisof the rms of a hundredof selectedaftershocks
(see The nodal plane striking 201ø and dipping45ø to the
belowconcerningaftershockslocations). The inversional- north westis in excellentagreementwith the observed
geo-
gorithm yields a strike of 201ø, a dip of 45ø, and a rake of metry of surfacefaulting describedin the previoussection
283ø (Figure9a) andthesethreeparameters
arequiteinde- and with the fault geometryoutlinedat depth by the af-
Fig. 7. (•) View to the northeastof the northernsegment of the fault fromEleochoriroad(site5 on
Figure2•). The limestonehill with the triangularfacetin the foregroundto the right marksthe uplifted
footwall. The ruins of Eleochori village lie on top of that hill. Note the unfaulted transverse range in the
background.(•) Break at site6 (seeFigure2•). View to the southwest.The scaleis givenby the pencil.
(½)Cracksacrossa dirt roadat the northerntip of the rupturezone(site8 on Figure2•). The vertical
componentof slip is • cm down-to-thenorthwest(to the right of the picture).
TABLE 1. Hypocentral Parameters
Latitude Longitude Depth, km ML Ms rnb
Main ShockSeptember15, 1986, 1724 GMT
Athens(NOA) 37ø 6.0' 22011.4 ' 1.0 5.5 6.4

CSEM 37 ø 7.8' 22ø12.0' 41.0
Thessaloniki* 37 ø 6.6' 22008.4 ' 8.0 5.8
NEIS 37ø 6.0' 22011.6 ' 11.0 5.8 6.0
AftershockSeptember15, 198t•, 1141 GMT
Athens(NOA) 37ø 4.8' 22004.2 ' 1.0 4.8 5.3

CSEM 37 ø 2.4' 22010.2 ' 37.0
Thessaloniki* 37 ø 2.4' 22007.8 ' 8.0 5.4
NEIS 36055.8 ' 22010.8 ' 10.0 4.8 4.9
* Papazachos
etal.[1988].
tershocks. The fault mechanism obtained here differs from Over a hundred of aftershocks per day were recorded,
that obtainedby Papazacho8et al. [1988]from first motion and 740 were located on the basis of having at least five
readings(strike=204ø, dip=51ø, rake=303 ø) and that of readingsof P arrivals(Figure 8b). Below,we first discuss
Dziewon8l•i
eta/. [1987](strike=211 ø) the choice of a velocity model and the precision on the
ø, dip=56ø, rake=304
essentialyby having a smaller componentof left-lateral slip. locations as well as the procedure used for focal mechanisms
determination. We then discuss the aftershocks distribution
Figure9b (row 7) showsa comparaison
of observed
andsyn-
and focal mechanisms based on the most reliable locations
thetic seismogramsfor the solution of Papazachoset al. SH
wavesat BUL, KBS, and KEV are almost nodal for this solu- and on individual focal mechanism solutions.
tion which disagreeswith observations. Moreover, polarities

Velocit•l Model and Location Uncertaintie•
read on long-period WWSSN recordsat stations STU and
IST which are compressionalconstrain the solution and were No detailed information on the velocity structure of the
misplottedby Papazachos et al. [1988]. At thesestations area is available. Thus, in order to find a reliable velocity
the first arrival is a Pn phase and assumingan upper mantle modelwe selected100 well-recorded aftershocks (with more
velocityof 8 km/s and a lowercrustvelocityof 6.5 km/s, than six P and two S readings)that wouldbe a representa-
the incidenceangleis about54ø insteadof 48ø as shownby tive sample of all the aftershocks,both in location and depth
Papazachoz et al. [1988]. Thusthe solutionobtainedagrees and first located these events in a half-space with a P velo-
with the first polaritiesread on long-periodrecords(Figure city between5.0 and 6.2 km/s, the Vp/V• ratio beingkept
constant and equal to 1.73. There is a clear minimum in
RECORDING AND ANALYSIS OF AFTERSHOCKS the mean rms of the 100 events for a P velocity between 5.5
and 5.4 km/s (Figure10). Moreover,lookingat the mean
A few days after the main shock, a team from the Ob- rms by depth range, the minimum occurs at a P velocity
servatoryof Athens and from Athens Universityinstalleda around5.3-5.4km/s for earthquakesshallowerthan 5 km.
portablenetworkof sevensmokedpaperinstruments. This In contrast, for earthquakes deeper than 5 km the minimum
network was complementedby nine stationsfrom the Insti- occursat a P velocitycloseto 5.7 km/s andfor earthquakes
tut de Physiquedu Globeof Paris and from the Observatoire deeperthan 7 km at a P velocitycloseto 5.8 km/s (Figure
de Grenoble. Thus, for a period of 10 days from Septem-
ber 18 through September 27, 16 smokedpaper instruments
(Sprengnether
MJ•Q 800) equippedwith 1-Hz verticalgeo-
phoneswere deployed. They provided a good coverageof
the aftershockarea (Figure 8) and allowedboth for well-
resolved hypocenters and reliable individual focal mecha-
nism determination. Station coordinates are listed in Table
2. All the instrumentsrecordedat 120 mm/min, yieldinga
time reading precisionof 0.05 s for P waves,except for some
far away stations like PET or KVAS that were recordingat
60 mm/min. In this casethe readingprecision
is only0.1 s.
S wave arrival times were also read on vertical components.
Often, at the close stations, S waves were impulsive, and
only in this case were they taken into account and always
weighted down with a factor of 2 or more compared to P
waves. For S waves, the estimated reading error is 0.2 s.
Every day (or every 2 daysfor thosestationsrecordingat
60 mm/min) the internalclockof the recorderwaschecked
21•o 22•o 22•o
against a radio signal in order to control the drift of the
internal clock which was usually between 0 and 0.1 s and Fig. 8a. Network geometry and outline of the aftershock
alwayslessthan 0.2 s/d. area.
I • I
37•o 1POL ,0 , 2 , 4Kin--

37•o
• •/••i•
(••NED
3•o•/ •- • o
ß M = 0.4
• M = 1.4
36• 36•
Fig. 8•. De•sil of •he sf•ershockdistributionwi•h t40 locs•edevents.Fsul•ssre from Figure 2• snd
•he hesvy line is •he obse•ed surfscerupture on •he Kslsms•s fsul•. Solidsqusressre stsfionlocsfions,
•he solid circle is •he locsfion of •he s•rong motion instrument.
220
I !
0 2 4 Km
37 8 37 8
37 3 -c/-. •_• • ¾j,• •. -'v _ 37 3
• M= 8.4
22
I 1
,I
22 o 2•z
3658
Fig. 8e. Map viewof the 344 bestlocatedaftershocks

(seetext for the selectioncriteria).a, b, c and
d are locations of crosssectionsin Figures 14 and 13. Large and small bold symbols indicate locations
of the September 13 mainshockand of the September 15 largest aftershock,respectively,given by the
CSEM (pluses),NOA (triangles)and Papazaehos
et al. [1987](crosses)
(Table1). Faultsandaftershock
locations shown as in Figure 8b.
KALAMATA 1986.09.1,5 to study the influenceof the Vp/V• ratio we foundthat it

had almost no effect on the mean rms at any P velocity.
Thus a Vp/• ratio of 1.73 waschosen. We then tried to
KM KJ find a layered velocity structure that would account for the
features described above. Because of the apparent strong
velocity gradient in the upper 5 km of the crust and in or-
DL
der to avoid too large velocity contrasts between layers for
which we did not find evidencesand that would create pro-
blems, especially when computing incidence anglesfor focal
WH
mechanisms,
weuseda three-layer
insteadof a simpler
two-
layer velocity structure. The final velocity structure chosen
(Table
2)yields
atalldepths
asmaller
mean
rms
than
any
of
the
half-space
model
(Figure
10).This
isprobably
inpart
dueto the largernumberof parameters
involved
but also
OF reflects the better adequacy of such a velocity structure at
$E
locating
shallow
events
structure.inparticular
than
just
ahalf-space
BK NI
U ", A Aftershocks were selected on the basis of the quality
of their location. The events were located using the
HYPO71 computerprogram[œeearidœa/•r,1975],which
gives standard errors ERZ, ERX, ERY for depth, latitude,
5's STF and longitude respectively. It is well known that those
3'Os errors underestimate the true errors on absolute location
[e.g.,//•t•eld •t •1., 1987]but still are usefulin estimating
the quality of a location. The following criteria were used:
KE KI rms<0.15 s, ERZ<I.2 kin, ERX•0.6 kin, ERYC0.6 km in
addition to requiring at least five P and one $ readings and
that the distance to the closeststation be smaller or equal
to the depth of the event for events deeper than 3 km. For
shallower events the last two conditions were changed to
WD requiring
at leasttwo$ readings
andthattheclosest
station
be at a distance lessthan twice the depth of the event. This
choice
wasmade
tocombine
theneeds
forprecise
location
and homogeneityin the aftershock distribution. Among
events thatareselected
than 0.1 s.
thisway,veryfewhave
armslarger
HA Because the network is dense and covers the aftershock
B area quite well, the locations are rather insensitive to the
BH NO
velocity structure. For all the modelstested (i.e'7 half-
U ',••
L ,• - space
the with
velocity
variation from
5to6
inlocations km/s,
given
bythe two
orthree
differentlayers)
models
isless
than 200 m for eventswithin the network(i.e., almostall
Fig. 9•. Main shockfocalmechanism
obtainedby simul-
eventsthat passedthe selectioncriterion).The variationin
taneousinversion
of P (top) and $/-/(bottom)long-perioddepth is in mostcasesnot larger than 1 km exceptfor events
WWSSNwaveforms. Observed (solidcurve)and synthetic shallowerthan 3 kin. Thoseare clearly better locatedwith a
(dashedcurve)seismograms are plottedscaledto the seis-
meanP velocitybetween5.4 and 5.0 km/s than 5.6 or more,
mic moment.STF is the sourcetime function;the ampli- but the precisedepth dependsstronglyupon the velocityof
tude scale is in microns. Parameters used are listed in Table
the upper layer and the calculateddepth couldvary from 0.5
4. Each seismogramis representedby a letter on the fo-
to 2.5 km quite easily. However, from these tests it is clear
cal sphere. Open circles show additional P wave dilatational
that eventswith a computed depth lessthan about 3 km
firstmotions
readonlong-period
records.
Solidcircles(com-
pressionalfirst motion) are stationsSTU and IST discussed and that meettheselection criteriadefined above(e.g.have
2 $ readings and the distance to the closeststation is smaller
in the text. Open squareis it'axis,blacksquareis P axis.
than twicethe depthof the event)arereallyshallower than
3kin. Thesetestsalsoindicatethat theerrors(ERZ,ERX,
ERY) are realisticestimatesof relative locationserrorsof
10). Thusthereis apparentlya strongvelocitygradientin the eventsone with respectto another,exceptfor the error
the upper 7 km of the crust. This resultis in agreementwith on the depth of events shallower than 3 km which could be
the interpretationof a refractionprofilepassingthroughthe 2 km or so.
plain of Kalamata (Kampos-Kalamata-Kyparissia; about 5 We present the location of the 344 selected aftershocks in
km southwest of the network[Z)•lCb•½••½al., 1987])that map view (Figure8), in crosssectionparallelto the fault
showsa 5-kin-thicklayerwith a 5.2 km/s P velocityovera trace(Figure13) and in crosssections
perpendicularto the
layer with a 6 km/s P velocity.In doingthe sameexercise fault trace(Figure14).
KEY P HYB P SLR P CHO SH BUL SH BEC SH KBS SH
Fig. 9b. Sensitivity of waveformsto variations in strike, rake and depth at sevenselectedstations.
Row 1 is the solutionin Figure 9a. Row 2-6, one parameterwasfixed (strikein rows2 and 3, rake in
rows4 and $, depth in row 6) and waveformswereinvertedfor the remainingones.The arrow on row
6 at KBS indicatesarrivalof s$ phase.Row 7 is Papazachos et al's [1988]solution.Numberson top of
eachfocalsphere
arel•strike/dlp/rake/depth/moment
x1016.
Focal Mechanisms TABLE 2. Station Coordinates and Velocity Structure

Station polarities for the events selected above and that
had more than six readings were plotted on a focal sphere Code Latitude Longitude
using incidence angles given by HYPO71. From these
plots, only the events for which nodal planes could be EVA 37ø06.93 ' 21ø57.82 '
drawn without too much uncertainty were kept and will be KAR 37ø05.05' 22 ø11.45'
KAT1 37ø04.24 ' 22 ø15.84'
discussed in this study. As pointedout before[e.g.,Ouyed
KAT2 37ø05.23 ' 22 ø15.47'
et al., 1983],the assumedvelocitystructuredoesnot have KDIO 37ø04.68' 21ø51.39'
much influence on the determination of the nodal planes for KEN 36056.59 ' 22ø15.37'
the deeper events. The main problem is with focal solutions KIT 36ø55.73 ' 22o09.05 '
that involve refracted rays. This occurs for events located KKAL 37ø03.16 ' 22ø08.24 '
KKAM 36 ø57.11' 22 ø11.48'
abovea velocitydiscontinuity
(here2 and5 km). For those KNED 37ø07.19' 22ø12.47'
rays the projection on the focal sphere may strongly depend KPER 37ø02.84' 22 ø10.85'
upon the assumed velocity structure. We choose to show KPRO 36ø52.71 ' 22 ø14.93'
only the solutions for which the orientation and nature of KVAS 36 ø46.88' 21 o55.83'
the focal solution remain roughly unchangedif no velocity PET 36ø54.26 ' 21ø51.31 '
POL 37ø09.66 ' 22ø07.87 '
discontinuity was present. Often the solutions have only RIZ 37ø01.24' 22 ø15.57'
one well-constrained nodal plane. In this casewe chooseto THO 37 ø07.05' 22 o04.53'
show the solution that minimizes the strike slip component
of the mechanism. P Velocity,km/s Depth, km
We calculated 133 focal solutionsand showtheir polarity
readings in Figure 11. All parameters for these solutions 5.0 0.0
are listed in Table 3. We also present map views of the 5.4 2.0
distributionof focalmechanisms by depthrange(Figure12) 6.0 5.0
andcrosssections perpendicular(N105ø) to the meanstrike

of the surfacefault trace(Figure14). Poisson's ratio is 0.25
0 all events
RMS
(S)lß ß depth< 5km
0'13
l ß depth>5
ß depth >7km
km
0.1
2
Il
0.1 e•e
[-I 3 layer model
o.,
oT
0.09 [--I
5!0 5!2 5!4 5!6 5!8 6!0 Vp(km/s)

Fig. 10. Plot of mean rms given by HYPO71 of 100 selected aftershocks versus the assumedhalf-space
P velocity. Note that the velocity at which the minimum rms occurs depends upon the depth of events
considered. The three layer velocity model chosenis listed in Table 2.
AFTERSHOCKS DISTRIBUTION within zone F, and the activity is continuous from about
AND FOCAL MECHANISMS 2 to 10 km. The remaining events occur below F, in the
footwall.
The aftershocks cover an area of about 15 by 10 km and
The seismicityin the southernpart of the cluster(Figure
can be separatedinto two clustersin map view (Figure 14½)is morediffus•andseemsto occurmostlyin the footwall
8). The main clusterlocatedto the southcan clearlybe as defined by the extension of zone F to the south. Such
associated with the observed surface rupture and will be
deformation in the footwall may be related to the fault bend
called below the Kalamata fault cluster. Farther north,
in the vicinity of the crosssection(Figure 8) wherethe
the second cluster is located away from the surface rupture,
surfacefault strike changesfrom about N30øE in the north
within the relay zone north of the Kalamata fault zone,
to N10øEin the south. Alternatively,the fault planemaybe
betweenthe Thouria and Kalamatafaults (Figures2a and steeper(___ 55ø-60 ø) therethanfarthernorth(whichwould
8b). Betweenthe two clustersthereis a well-defined "gap" not be surprising in view of the topography and morphology
of aftershock activity over a width of 2 km and down to
of the two areas).
a depth of about 8 km (Figures8 and 13). Thesetwo
separated zones of seismicity seem to join at a depth of 9-10
Almost all focalmechanisms within zoneF (both in the
northern and southern part of the cluster; Figures 12, 14b,
km (Figure13). No eventsdeeperthan 12 km werelocated,
most of them being confined to the 10 first kilometers of the and 14½)shownormal faulting consistentwith the main
crust.
shock mechanism, continuouslyfrom a depth of 10 to 4 km.
However, the dip of the individual fault planesof aftershocks
Kalamata Fault Cluster
mechanism
are steeper(55ø or 60ø) than that definedby
both the aftershocks distribution and the main shock focal
mechanism
On crosssectionsb and c perpendii:ular to the fault trace (45ø). Fortheseevents(e.g.,1, 5, 8, 35, 44, 50,
(Figures14band 14½)the clusterappearsclearlyas a nor- 75, 119, 121) the dip of the fault planeis usuallyconstrai-
thwestwarddipping zone, although in its southernmostpart ned by the polarity reading at KPER. In most casesa clear
(Figure 14½)the seismicityseemsto be morediffusethan compression was observedat this station which liesjust on
in its northernpart (Figure14fi). On Figure14fiwe show the fault trace. Without a very reliablevelocitymodelit is
the observedsurfacegeometry with the fault plane dipping not easy to evaluate the significanceof such an observation.
about 50ø as well as its downwardextension,dipping45ø, Variationsof about10ø of the incidence anglecanprobably
in agreement with the main shock focal mechanism solution occurby perturbingthe velocitystructure(i.e., by introdu-
(i.e., the "average"fault planeshownwith a dashedline on cingvelocitygradients)or because of inhomogeneities
in the
Figure 14fi,Table4). The two parallellinesbounda zone vicinity of the fault plane. Thus the discrepancybetween
labeled F within which all hypocentersare at a distanceless the fault planedip of aftershocks
andthat of the mainShock
than 0.85 km from that "average" fault plane. This dis- could be a consequence of the simplestructure used. Ho-
tance correspondsto a variation in depth of + 2 km, which wever,if this discrepancyis real, it can be comparedto the
representsthe depth location uncertainty. Thus all hypo- steepenichelonfault planesobserved
at the surface(• 60ø-
centersenclosedin zoneF can be considered,within location 70ø) asopposedto a lesssteepaverage
fault planededuced
uncertainties, to lie on the master fault plane defined both fromdetailedmapping(__50ø) (seeFigure6). Suchan en
by surfaceand deep data. Clearly, the great majority of ichelon geometry of faulting has often been observedat va-
eventsin the northernpart of the cluster(Figure 14b) lie riousscalesandwithindifferentrocktypes[e.g.,Tehalenko,
• ? 8 D
1• 17 18
)
21 22 2• 2• 25
2.8 2.7 2.8 2.9 :30
36 37 38 39
Fig. 11. Focalmechanisms of 133 aftershocksplottedon the lowerhemisphere. Solidcirclesare

compressions;opencirclesaredilatations.Refractedarrivalsareindicatedwith largercircles.P and T
axisarealsoindicated
in additionto theslipdirection
oneachof thenodalplanes.Earthquake andfocal
solution parameters are listed in Table 3.
4G 47 48 49 50
51 52 5• 54 55
71 72 73 74 75
78 77 78 79 80
81 82 8• 84
Fig. 11. (continued)

101 102 10S 104 105
108 107 108 10• 110
1 1 1
126 !27 128 129 1SO
1S1 1S2

TABLE 3. List of Location Parameters for the 133 Aftershocks Shown With their Focal Mechanisms in Figure 11
Day Time Latitude Longitude Depth, Mag- rms, ERX, ERY, ERZ, Strike, Dip, Rake,
Event Sep.1986 GMT 37ø 22ø km nitude s km km km deg deg deg
I 18 1452:23.5 3.98 7.76 4.73 1.84 0.12 0.4 0.4 0.6 190. 50. -97.
2 18 1527:01.4 7.06 9.13 3.39 2.23 0.14 0.4 0.5 1.1 205. 75. -107.
3 18 1614:55.2 4.46 7.76 4.85 1.42 0.15 0.6 0.5 0.5 180. 45. -90.
4 19 0146:30.3 6.86 7.82 9.43 2.40 0.04 0.2 0.2 0.5 190. 45. -90.
5 19 0229:32.3 3.45 7.53 4.14 0.89 0.07 0.6 0.3 0.5 180. 42. -90.
6 19 0309:47.9 1.44 9.14 2.72 0.99 0.07 0.4 0.3 0.9 158. 50. -91.
7 19 0414:06.4 6.35 7.24 9.15 0.99 0.07 0.4 0.6 1.0 193. 45. -92.
8 19 0743:00.2 3.47 8.67 4.40 2.07 0.07 0.3 0.2 0.3 190. 60. -90.
9 19 0824:06.8 2.89 8.61 5.00 2.14 0.04 0.1 0.1 0.2 175. 55. -96.
10 19 0942:18.3 7.01 9.55 5.73 2.02 0.07 0.2 0.3 0.6 220. 68. -68.
11 19 1007:02.8 3.95 9.42 5.00 1.97 0.10 0.4 0.4 0.5 108. 82. 96.
12 19 1053:13.3 0.81 9.11 5.44 2.06 0.08 0.3 0.2 0.6 100. 70. -90.
13 19 1207:25.6 7.17 9.67 6.12 2.17 0.06 0.2 0.2 0.5 220. 68. -66.
14 19 1208:17.5 7.40 9.85 5.88 1.89 0.04 0.1 0.2 0.4 220. 70. -68.
15 19 1211:37.3 7.41 10.04 6.70 1.85 0.04 0.2 0.2 0.3 220. 75. -73.
16 19 1404:50.2 3.66 10.53 5.00 1.67 0.07 0.3 0.3 0.5 205. 75. 145.
17 19 1411:36.0 0.65 8.98 4.18 1.59 0.08 0.4 0.2 1.0 199. 50. -127.
18 19 1449:43.6 2.72 7.81 5.00 1.81 0.08 0.2 0.3 0.4 218. 52. -91.
19 19 1732:28.8 4.04 10.64 3.52 1.43 0.04 0.2 0.2 0.4 200. 70. 129.
20 19 2056:49.3 6.81 9.34 5.55 1.98 0.05 0.1 0.2 0.4 220. 65. -51.
21 19 2322:34.9 3.70 9.53 5.22 1.71 0.09 0.2 0.3 0.6 125. 45. -146.
22 20 0006:18.6 6.62 9.50 4.10 1.77 0.07 0.2 0.3 0.6 215. 60. -71.
23 20 0006:39.5 5.31 11.05 5.20 2.04 0.10 0.3 0.4 0.9 172. 80. 85.
24 20 0147:04.9 3.63 9.46 3.31 1.98 0.10 0.3 0.3 0.7 131. 40. -122.
25 20 0235:52.7 6.69 9.76 6.03 1.65 0.08 0.4 0.3 0.6 358. 70. -33.
26 20 0456:06.1 6.99 9.47 5.90 1.64 0.08 0.5 0.5 0.7 206. 60. -42.
27 20 0542:32.6 4.33 6.71 6.91 2.35 0.15 0.5 0.6 1.0 290. 70. O.
28 20 0602:59.0 6.82 9.70 6.58 2.21 0.08 0.2 0.3 0.7 200. 65. -67.
29 20 0621:15.0 6.29 9.74 8.23 3.00 0.05 0.3 0.7 1.1 200. 65. -67.
30 20 0717:33.0 6.94 9.95 6.97 2.11 0.07 0.2 0.4 0.9 200. 65. -90.
31 20 1058:01.8 5.45 6.24 10.53 0.61 0.05 0.5 0.6 0.7 193. 60. -78.
32 20 1443:27.0 0.98 12.24 4.42 1.04 0.06 0.2 0.2 0.4 340. 90. 180.
33 20 1508:13.6 7.29 9.50 5.46 1.29 0.07 0.2 0.3 0.7 170. 25. -87.
34 20 1633:29.2 1.79 8.06 7.33 1.03 0.15 0.5 0.4 0.9 157. 65. -163.
35 20 1636:28.0 4.80 7.39 6.37 1.45 0.15 0.4 0.4 1.1 185. 65. -90.
36 20 1823:44.6 7.07 9.48 5.80 0.55 0.07 0.3 0.3 0.7 220. 70. -59.
37 20 1826:14.3 6.26 5.98 9.65 1.09 0.08 0.3 0.4 0.6 192. 48. -80.
38 20 1837:58.2 7.64 9.75 7.15 2.04 0.06 0.2 0.3 0.5 220. 80. -73.
39 20 1900:44.1 7.49 9.41 6.00 1.31 0.09 0.2 0.3 0.6 220. 65. -51.
40 20 1924:58.1 6.45 8.61 7.52 1.70 0.10 0.3 0.3 0.6 150. 80. 153.
41 20 2021:25.6 6.89 8.38 6.11 1.20 0.11 0.3 0.3 0.6 218. 65. -35.
42 20 2058:22.1 7.50 9.04 7.26 1.28 0.07 0.3 0.4 0.6 220. 65. -51.
43 20 2234:16.2 7.68 7.90 7.09 1.15 0.06 0.2 0.3 0.5 203. 35. -67.
44 20 2346:14.4 4.47 7.41 6.86 1.53 0.13 0.4 0.4 1.1 200. 60. -75.
45 21 0015:15.8 3.19 7.22 6.29 1.01 0.12 0.4 0.4 0.9 158. 60. -114.
46 21 0042:35.9 6.66 8.37 7.38 1.70 0.12 0.4 0.5 1.1 160. 60. -99.
47 21 0048:45.3 7.37 8.13 8.12 1.79 0.09 0.3 0.4 0.7 200. 60. -82.
48 21 0121:30.8 7.29 8.73 6.75 1.25 0.10 0.4 0.7 1.1 175. 55. -93.
49 21 0549:22.6 6.05 9.30 6.14 1.88 0.12 0.4 0.3 0.8 260. 60. -59.
50 21 0615:54.9 3.99 7.79 5.55 1.48 0.13 0.4 0.3 0.7 197. 60. -90.
51 21 1157:06.8 5.18 6.39 8.37 1.29 0.10 0.3 0.3 0.7 180. 65. -89.
52 21 1220:33.6 7.64 7.85 6.10 1.50 0.07 0.2 0.3 0.7 345. 60. -71.
53 21 1519:58.3 3.71 10.93 6.13 0.96 0.07 0.3 0.2 0.6 162. 45. 87.
54 21 1824:32.0 7.63 7.78 6.85 1.10 0.10 0.3 0.4 0.7 200. 35. -57.
55 21 1955:25.1 6.26 8.58 4.91 1.76 0.12 0.2 0.3 0.5 167. 70. -88.
56 21 2321:11.5 2.10 11.74 4.68 1.17 0.06 0.2 0.2 0.3 210. 40. 98.
57 22 0021:29.7 3.61 7.51 6.56 1.57 0.07 0.3 0.2 0.7 150. 60. -131.
58 22 0419:10.6 1.30 8.62 8.43 1.30 0.06 0.3 0.2 0.8 205. 60. -108.
59 22 0548:23.1 8.04 8.33 10.19 1.99 0.04 0.3 0.2 0.5 159. 50. -127.
60 22 0603:05.6 6.63 7.66 6.63 1.36 0.14 0.4 0.5 1.0 295. 60. -85.
61 22 1009:05.4 7.45 7.93 7.31 2.12 0.11 0.4 0.6 1.0 355. 60. -79.
62 22 1033:52.0 6.45 5.65 9.13 1.37 0.06 0.3 0.3 0.3 192. 63. -84.
63 22 1043:27.8 6.74 8.71 4.91 1.16 0.07 0.2 0.3 0.6 199. 50. -62.
64 22 1111:47.2 4.60 7.62 5.84 1.39 0.13 0.5 0.5 1.0 187. 55. -90.
65 22 1112:35.5 6.66 8.27 6.12 1.60 0.07 0.2 0.3 0.7 220. 65. -54.
66 22 1138:07.0 6.36 6.15 9.07 1.62 0.10 0.4 0.4 0.6 220. 70. -74.
67 22 1215:46.0 5.41 9.29 7.50 1.94 0.08 0.3 0.2 0.6 205. 80. 95.
68 22 1226:30.6 5.76 6.36 10.65 1.42 0.09 0.4 0.3 0.5 145. 70. -102.
TABLE 3. (continued)
Day Time Latitude Longitude Depth, Mag- rms, ERX, ERY, ERZ, Strike, Dip, Rake,
Event Sep.1986 GMT 37ø 22ø km nitude s km km km deg deg deg
69 22 1304:17.2 1.24 11.54 4.71 1.50 0.06 0.2 0.1 0.3 215. 35. 107.
70 22 1354:12.5 3.86 11.47 6.08 1.24 0.08 0.3 0.2 0.4 210. 65. -126.
71 22 1611:30.9 4.25 7.28 6.11 1.88 0.15 0.6 0.5 1.2 210. 53. -90.
72 22 1732:10.9 6.83 8.47 6.05 1.19 0.07 0.3 0.3 0.7 170. 45. -113.
73 22 1737:29.1 0.08 12.31 4.33 1.59 0.13 0.3 0.2 0.8 186. 60. 54.
74 22 1855:17.4 3.47 7.90 6.40 1.74 0.09 0.3 0.2 0.6 195. 60. -103.
75 22 1908:48.7 4.80 6.49 8.38 1.71 0.09 0.3 0.3 0.5 205. 75. -83.
76 22 2202:32.0 4.89 4.63 10.41 1.30 0.15 0.8 0.7 1.1 195. 50. -80.
77 22 2244:50.9 7.89 7.76 8.20 1.09 0.08 0.4 0.4 0.7 335. 75. -86.
78 22 2333:25.0 7.13 8.28 7.04 1.52 0.10 0.3 0.3 0.6 355. 70. -77.
79 23 0010:12.9 7.47 9.27 4.67 1.15 0.11 0.3 0.3 0.6 195. 60. -54.
80 23 0036:09.6 4.84 6.78 7.84 1.38 0.08 0.3 0.3 0.6 175. 60. -90.
81 23 0115:41.5 6.99 7.81 6.41 2.16 0.08 0.2 0.2 0.7 345. 60. -67.
82 23 0219:54.4 6.92 6.52 9.20 1.50 0.10 0.3 0.3 0.5 225. 70. 121.
83 23 0516:4•.7 4.42 7.28 5.31 0.84 0.05 0.3 0.2 0.5 190. 60. -90.
84 23 0608:23.3 1.60 8.90 3.79 1.33 0.11 0.5 0.2 0.8 193. 65. -90.
85 23 0913:10.1 6.91 6.94 8.84 1.17 0.08 0.3 0.3 0.5 103. 25. -90.
86 23 0914:13.8 7.19 7.13 8.45 1.17 0.05 0.3 0.3 0.5 103. 25. -90.
87 23 1122:10.9 1.15 10.78 5.25 1.74 0.09 0.3 0.2 0.7 70. 61. 43.
88 23 1136:22.4 1.11 12.77 6.70 1.21 0.09 0.3 0.1 0.5 39. 40. 116.
89 23 1352:05.5 7.30 8.57 7.83 1.42 0.07 0.2 0.2 0.5 350. 60. -74.
90 23 1412:16.7 7.26 8.10 7.91 1.59 0.13 0.4 0.4 0.7 355. 60. -71.
91 23 1449:29.9 1.73 10.60 6.54 1.72 0.13 0.5 0.3 1.1 185. 5. 90.
92 23 1510:43.2 7.13 8.26 7.18 1.34 0.06 0.2 0.2 0.3 340. 60. -79.
93 23 1613:33.1 7.82 8.28 7.92 1.11 0.04 0.1 0.2 0.3 180. 60. -77.
94 23 1655:24.8 8.44 •.43 8.02 0.99 0.06 0.3 0.4 0.6 335. 40. -90.
95 23 1727:06.6 7.29 7.44 7.12 1.37 0.11 0.3 0.4 0.6 155. 60. -93.
96 23 1731:40.4
7.87 8.01 9.64 1.49 0.07 0.3 0.4 0.6 155. 65. -92.
97 23 1750:13.4 7.94 7.76 7.63 1.20 0.09 0.3 0.5 0.7 155. 60. -93.
98 23 1753:39.2 7.65 7.53 7.72 1.36 0.10 0.3 0.4 0.6 155. 40. -110.
99 23 1829:30.2 8.16 8.11 7.70 1.55 0.05 0.3 0.3 0.7 155. 60. -93.
100 23 1928:55.5 7.73 7.12 8.72 1.12 0.10 0.3 0.5 0.7 155. 60. -93.
101 23 1931:22.2 7.99 7.75 7.73 1.23 0.11 0.3 0.5 0.7 155. 60. -93.
102 23 1935:17.3 7.58 7.22 õ.04 1.38 0.17 0.5 0.7 1.1 155. 60. -93.
103 23 2009:25.4 8.03 7.60 8.61 1.17 0.05 0.2 0.2 0.3 155. 60. -93.
i04 23 2325:33.5 3.81 8.17 6.77 0.98 0.07 0.2 0.2 0.4 185. 60. -90.
105 24 0036:46.0 8.15 9.21 3.84 1.11 0.13 0.3 0.4 1.1 200. 52. -77.
106 24 0037:51.7 7.87 9.•2 3.59 1.17 0.06 0.2 0.3 0.7 200. 52. -77.
107 24 0208:21.7 8.15 7.91 11.20 1.86 0.07 0.3 0.3 0.5 195. 35. -50.
108 24 0228:45.8 7.86 8.32 8.63 1.14 0.08 0.2 0.3 0.4 325. 60. -87.
109 24 023,2:24.0 7.62 8.96 2.92 1.16 0.06 0.1 0.2 0.6 200. 52. -77.
110 24 0256:21.0 2.61 11.25 5.00 0.82 0.09 0.3 0.2 0.5 245. 80. 97.
111 24 0345:54.2 7.27 7.97 6.85 1.13 0.10 0.3 0.4 0.7 325. 60. -87.
•12 24 0444:42.6 7.17 7.31 8.85 0.93 0.14 0.5 0.5 1.0 103. 25. -90.
113 24 0448:49.3 7.10 6.81 10.20 1.22 0.08 0.4 0.4 0.6 103. 25. -90.
114 24 0449:56.4 7.44 7.28 8.85 1.18 0.10 0.4 0.5 0.8 103. 25. -90.
115 24 @715:07.6 7.70 7.53 7.93 1.52 0.11 0.3 0.4 0.7 180. 40. -90.
116 24 1856:33.9 7.32 7.65 9.22 1.88 0.07 0.3 0.3 0.5 234. 30. -50.
117 24 2024:50.2 1.95 7.37 8.04 2.11 0.09 0.4 0.3 1.1 155. 56. -68.
118 25 0037:49.2 4.38 7.75 6.69 1.67 0.12 0.4 0.4 1.1 200. 40. -106.
119 25 0332:22.2 4.72 6.91 8.69 1.73 0.11 0.4 0.5 0.9 185. 60. -90.
120 25 0552:30.2 4.42 6.75 6.97 2.07 0.08 0.4 0.4 1.0 190. 60. -90.
12! 26 1108:14.3 4.41 6.53 7.94 1.50 0.08 0.3 0.3 0.4 185. 60. -90.
122 26 1238:06.4 4.•5 6.78 8.10 1.26 0.11 0.4 0.4 0.6 180. 60. -90.
123 26 130.7:45.9 8.27 7.95 9.30 1.45 0.12 0.7 0.5 0.9 103. 38. -90.
124 26 1409:53.9 7.68 7.34 7.31 1.11 0.10 0.5 0.6 1.1 140. 58. -114.
125 26 1528:08.1 8.46 7.47 7.40 1.36 0.15 0.5 0.6 1.0 285. 40. -90.
126 26 1937:21.1 1.72 10.i0 3.20 1.31 0.15 0.5 0.3 1.1 305. 80. 88.
127 26 2142:45.8 6.83 8.43 4.87 0.97 0.11 0.3 0.3 0.7 185. 50. -90.
128 26 2236:26.1 6.42 7.07 9.04 1.86 0.09 0.3 0.3 0.7 90. 2. -23.
129 27 0105:03.8 3.48 7.74 6.55 1.14 0.14 0.4 0.3 0.8 205. 60. -67.
130 27 0132:20.3 5.52 5.76 9.11 1.30 0.09 0.3 0.3 0.4 180. 50. -90.
131 27 0149:19.0 0.54 7.52 7.15 1.38 0.15 0.6 0.3 1.1 195. 67. -107.
132 27 0328:33.6 0.35 7.08 7.02 0.96 0.15 0.6 0.4 1.0 195. 67. - 107.
133 27 0635:32.9 2.88 8.40 5.18 1.81 0.10 0.2 0.2 0.5 195. 67. -110.
22 2 22 7 2212 2217
I 't It' I
371o 371o
375 37 5
•) KAT 1
0 KKAL (•/• • 11o
37 o 37 o
73
/ oKEN
/
IT
3655 3655
22 2 22 7 2212 2217
Fig. 12. Map views of 133 focal mechanisms

plotted on the lower hemisphere(solid quadrantsare
compressional
ones)by depthrange:(a) 0-5 km, (b) 5-7 km, (½) 7-9 km, (d} 9-12 km. Earthquakeand
focal solution parameters are listed in Table 3. Individual focal solutionsare shown in Figure 11 with
polarity readings.
1970; Granier,1985]. The differencein anglebetweenmi- gnificantly from that of the main shock, with even some
nor en 6chelonfault planes and the main fault plane is also pure thrust faulting mechanismswith P axis oriented about
wenknownin laboratoryexperiments [e.g.,JaegerandCook., Nl10ø:h20ø (events16,19,53, 69, 73, and 88). This thrust
1979]. Theseen6chelon planesrepresentinitialshearfailure faulting is representedby a small number of events but ac-
underfrictional regime. Then the 60ø dippingfault planes counts for most of the postseismic deformation in the foot-
of aftershocksreported here may also representdeformation wall Thrust faulting in the footwall of a normal fault has
on such an en 6chelonfeature, implying that the aftershocks already been observedin the Wasatchfault area in Utah
broke patchesnearly on the main fault plane. Alternatively, [Arabasz
et al., 1980]. ZandtandOwens
[1980]developed
those aftershocks mechanismscould represent slip on anti- an elasticflexural plate model to explain suchobservations.
thetic planes dipping to the east, which would balance the However,postseismic stressreleasein the compressedupper
internal rotation in the vicinity of the main fault plane. part of the footwallappearshereas a morestraightforward
All focal mechanisms of events in zone F in the northern interpretation.
part of the cluster are consistentwith an horizontal T
Northern Olu•ter
axis oriented N105ø:i:15ø and an almost vertical P axi•
and are compatibleboth with the main shockand surface This cluster appear• to be more localized in map view
observations. (Figure8) as well as in crosssection(Figure14a)than
differ si- the one to the south.
Only in the footwall do the focal mechanisms Although no surface fault trace
22 2 22 7 2212 2217
3710
375
370
3655
2• 2 22 7 2212 2217
is associatedwith this cluster, one can project, at depth, than in the southern one, and some are lesswell constrained,
zone F into this area (Figures 14a and 3a). Between but they still can be classified into three sets.
depths of 10 and 4 km, although the seismicity seemsto A first set contains mechanisms for which the T axis
be shifted, the projection of zone F could possiblyrepresent
azimuthis consistent with that of the mainshock(NllSø).
the northern extensionof the master fault plane. However, It includesevents(e.g., 10, 13, 14, 15, 20, 26, 28, 30, 36, 38,
at shallower depth zone F cannot be defined because of 39, and 42) occurringnear the largestaftershockrecorded
the lack of seismicity. There is a concentrationof large by the temporary array on September20 (29). They
aftershocks (Mz,;•3.0) at the top of this seismiczoneat a are characterizedby normal faulting on a steepplane ('•'
depth of 4-6 km, the largest aftershockrecorded during the 70ø) orientedapproximately N220ø. Thisset alsoincludes
field experiment(September20 at 0621 GMT, 37ø6.94'N, shallowereventsoccurringwithin or abovezoneF (e.g. 22,
22ø9.95'E,M•.-3.4) alsotookplacehere(Figure14•). The $5,63, 79, 105, 106, 109, and 127). For all theseeventsthe
absenceof shallow seismicity is real and is not an artifact of T axis is close to the horizontal and oriented N105ø:!:15 ø.
the networkgeometrybecauseoneof our stations(KNED) The second set contains mechanisms for which the T axis
lies very close to this zone and thus constrains shallow azimuthis rotatedabout45ø (aboutN70ø) on the average
events nearby. In any case, upward propagation of the with respect to that of the main shock. It includes events
hypothetical fault plane F could be somehowhindered by that have been drawn as almost pure normal fault with a
the nappe structure at a depth of 4-6 km, which may be T axisorientedN70ø:E10
ø (e.g.,52, 61, 77, 78, 81, 89, 90,
similar to that depicted in Figure 3• at a depth of 1-2 km. 92, 96, 98, 102,107,and 124). Asidethe better constrained
Focal mechanisms in the northern cluster are more varied mechanisms (96, 98, 102, and 124 and probablyalsoevents
371 371o
10 Km
KNED
90
(• KATZ
37 37 5
0 KAT !
40
/•] RI Z
370 --370
KEN
3655 , 3655
22 a 22 7 2212 22i7
94, 97, 99, 100, 101, and 103), the othermechanisms

could DISCUSSION
also represent almost pure strike-slip solutions with right-
lateral slip on a steep plane oriented more or less east-west Rupture Dimensions and Displacement
and thus perpendicularto the Kalamata fault (seefocal The obeerved coseismic displacement on the fault trace
solutionsin Figure 11). In fact, almostany intermediate is between 6 and 18 cm on the average. Taking a mean
solution from a pure strike slip to a pure normal fault displacement of 15 cm, 8 km for the average fault length
solution would be consistent with the data, in all casesthe (2 km more than the surfacebreak), 14 km for the fault
strike of the T axis would remain about N70ø-4-15 ø.
width(alongdip) and an elasticmodulus p =3.3x10lø
The third set contains mechanisms for which the T axis N/m2, weobtaina seismic
moment of5.5x10l? N m which
azimuth is rotated about 90ø with respectto that of the is within the error bound of the value of 7.0+2.5x10 •? N
main shock. It includes only small events located near m obtained from waveform modeling. A seismic moment
the southernedgeof the northerncluster(exceptevent125 of 7.0x10
l? N • yields
a meandisplacement
of 18.5cm
locatedfarther north). Theseeventsgenerallyhave one for the fault dimensions listed above. This may indicate
well-constrainednodal plane orientedN100ø and dipping that the observeddisplacement at the surfaceis smaller than
northabout80ø (60, 85, 86, 112, 114, 125, 113,and123). the mean displacement on the fault due to stress relaxation
The other plane is not well constrained,and we represent near the surface. Alternatively, if the whole surface of
here the mechanismswithout any strike-slipcomponent. In the aftershock
zone,includingfloatof the •gap•, is also
any casethe mechanismshave a T axis oriented N20ø+10 ø. taken into account to evaluate the area ruptured by the
2217
I I I
82 113 102 123 __ 107
371o ..
3710
10 Km
I -• 9o(.•
KNED
O KAT2
375 375
O KAT1
('•
KKAL
(•(•ER
• RIZ
370 37 o
KEN
/
d
3655 3655
main shock, the estimate of the seismic moment would be distribution of the Kalamata fault cluster. Within this
increasedby about 30%. In any case,the resolutionon the cluster most of the aftershocks occurred either on the fault
value of the seismic moment does not allow to differentiate plane or in the footwall of the fault. Eventson the fault
between these two hypotheses concerning the extent of the plane have focal mechanismsshowingnorthwest-southeast
ruptured area. extension in agreement with the main shock.
Such a consistencyis unusual in well-documented dip-
Kalamata Fault
slip earthquake sequences,even those with clear surface
The most striking result obtained in this study is the very faulting. Often aftershocksdo not define a fault plane
goodconsistency betweensurfacedata (i.e., geometryand (e.g.,Thessaloniki [Soufleris
et al., 1982],Corinth[Kinget
kinematics of both coseismicsurface faulting and long term •., 1985],andE1Asnam[OuFe4et ,•., 1983]). In addition,
quaternaryfaulting) and seismological data at depth (i.e., aftershockmechanismscannot generally be easily integrated
aftershock distribution, main shock focal mechanism and in a singlefaulting model (e.g., Irpinia
individual aftershocksmechanisms). They all point to a Kin•, 1984])evenwhenaftershocklocationsdefinea main
simple normal fault geometry with a fault plane striking fault planeasfor the BorahPeakearthquake[Richinset al.,
about N15ø and dipping about 45ø northwestward. The 1987]. All the studiescitedaboveconcern earthquakes with
observed surface faulting length is 6 km, but the fault magnitude 6.4 or more, with aftershockzone at least 3 or 4
plane is continuous over a length of about 8 km from the times larger than that of Kalamata but coveredby about the
subsurface to about 10 km as shown by the aftershock same number of stations. The minimum distance between
N15=E to the presence of a large number of aftershocks and to a

KNED complicated pattern of focal mechanisms. For moderate-
I size earthquakes, with magnitude between about 5.5 and
7, any intermediate situation may exist. The dimensions
of the rupture plane along strike and dip are comparable
(aspectratio _ 1) and the rupturemay or may not reach
the surfacedependingupon local conditions(Figure 15).
For instance,in Irpinia (Ms--6.9), the rupturereachedthe
(DO
surface only in few places, representingless than 20% of
its total length[ Westaway
andJackson,1987;Bernardand
Zollo,1988]and the focalmechanisms of aftershocks
were
highly varied IDeschamps and King, 1984]. In contrast,
in Kalamata (Ms--5.8), the surfacerupturewasclearand
coveredabout 2/3 of the aftershockzonelength. In both
instances, normal faults cut acrossa nappe structure made
0
o
of platform sedimentary rocks, but in Irpinia the presenceof
a thick pile of fiysh under a massive rigid layer of limestones
12
may hinder rupture propagation to the surface.
Finally, an important and rather uniqueobservation(to
our knowledge) is the natureof focalmechanisms
of events
on the fault plane versus those of events in the footwall.
Events on the fault plane all have focal mechanisms indica-
Fig. 13. Hypocenters projected on crosssection oriented ting northwest-southeast extension like the main shock. In
NI$øE, parallelto the averagestrike of the Kalamata fault contrast in the footwall focal mechanismsusually showclear
(d in Figure8½). northwest-southeastcompression. This compressioncan be
understood qualitatively as the result of postseismicstress
release in the compressedupper part of the footwall. The
stations is very rarely less than 10 km • opposed to this comparable
observation
madeby Arabasz
eta/. [1980]in the
study where, within the aftershock zone, stations are $ km Wasatch fault footwall was not directly linked to the occur-
or less apart. Thus the resolution on aftershock locations rence of a large earthquake like here. It could nevertheless
for this study is probably better than in most of the studies well represent the same kind of residual deformation rather
mentioned. This may contribute to the fairly simple picture than being associatedwith elastic flexure as proposedby
given by the aftershocks. However, for the 1980 earthquake Zandtand Owens[1980].
in Norcia (Central Apennines,Italy) whichwassimilarto
Main ShockLocation and Rupture Process
that of Kalamata both in size and aftershockzone coverage
but which did not ruptured the surface,there was no simple On the basis of a rough interpretation of strong motion
relationship between the aftershock geometry and the main records, we concluded before that the distance between Ka-
shockfocalmechanism
[De•champ•et al., 1984]. lamata city and the main shock epicenter was at least about
Although scaling laws relate earthquake seismicmoments 10 km unless a low-velocity sedimentary layer exists beneath
(or magnitudes) to their rupturedimensions[e.g.,Kanamori the city,in whichcasethe distancecouldbe smaller(but lar-
andAnderson,1975],smalland largeearthquakes mustbe ger than about7.5 km). Thusthe main shockepicenteris
distinguishedbecauseone rupture dimensionis fixed by the probablywithin the northern cluster (corresponding to a
thickness oftheseismogenic 1982]whichranges 8.5- to 12-kmdistance). This caseis illustratedby scenario
layer[Scholz,
from about 10 to 20 km (for crustalevents)dependingon I on Figure 16 where the rupture is assumedto have been
the regionalcontext[ChenandMolnar,1983;Meissnerand initiated in the northern cluster. Then the northern cluster
St•ehlau,1982]. Small earthquakes,with magnitudeless would be the northernmost extension of the Kalamata fault
than about 5.5, usually correspond to rupture dimensions plane and the •gap• of aftershock activity might represent
lessthan about10x10km2 with a meandisplacement
less an unbroken barrier on this plane. The other possibility
than 10 cm and may hardly ever reachthe surface(Figure is that the main shock lies a few kilometers to the south of
15). At the other extreme,earthquakes with magnitude the northerncluster(corresponding to a distanceof 7.5-8.5
larger than about 7 correspond to rupture areas greater km for the strongmotionrecord),withinthis•gap• of after-
thanabout500km2 andto meandisplacements
greater
than shocks. The BorahPeakearthquake[Richinset al., 1987]
about 1 m. Those earthquakes often produce surfacebreaks, is a well-documented example of such a situation where no
probably because the vertical dimension of their rupture aftershock activity has been recorded in the surroundingsof
plane exceedsthe thickness of the seimogeniclayer. For the main shock hypocenter. The main shock being located
those events, except possibly for low-angle thrusting, the within the gap, two possibilitiesstill remain. The first one,
rupture plane must therefore extend horizontally in length: scenario
IIa (Figure16) assumes
asscenario
I that thenor-
the shapeof the ruptureplaneis elongated(aspectratio > 1) thern cluster is the northernmost extensionof the fault plane
and the large slip per unit length to be absorbed implies and that the earthquake ruptured the entire aftershockzone.
ScenarioIIb (Figure16) assumes
a wide perturbed zone where strain energy is released, on insteadthat the rupture
the two sidesof the fault and aroundthe edges(Figure propagated only to the south and that the northern cluster
15). Thus aftershocks will spreadon both sidesof the is a diffuse zone of deformation.
fault and will not outline its geometry. In addition, the The complexity in the first 3-4 s of the rupture inferred
stress field around edges may also be perturbed, leading from a preliminary analysisof short-periodrecordssuggests
sw sw
I I I I I I I
o oO
o
0 (•
8 =_
29
Kalamata Fault Kalamata Fault

sw
$w
I I
/
/
I I I //2 I
o o
4 ,.,
0 0 •0 0
•/% 0 ø¸
12
12
$w
Kalamata Fault Kalamata Fault
o
I I I !
/
•'ø• .,/oø? ø
/0 o 0/u0 Or"-b o oO - . ..,.,
/ 8/o oø c•
/,2y.'ø'.,,
0o• o
FIB;.
14. Hypocentere
(left)andfocal
mechanisms
(ris;ht)
projected
oncrosssections
oriented
N105øE,
perpendicular
totheaverage
strikeoftheKalamata
fault(a,bandcinFigure8½).
Thefronthemisphere
projections
offocalspheres
areplotted.Because
ofspace problems,
onlya representative
sample
ofthe
focal
mechanisms
inFigure
13have
been
projected.
Numbers
refer
toTable
3andtoFigure
13.(•) The
northern
cluster
onFigure
8. (•) and(½)TheKalamata
faultcluster.
OnFigure14•theKalamata
fault
isrepresentedtodip50ønearthesurfaceand45øatdepth (dashed line).Thetwoparallel
dashedlines
define
zone F which encloses
allepicenters
thatcanbeconsidered
tolieonthefaultplane,
withinlocation
uncertainties
(see
text).OnFigures 14•and14½ theprojection
ofzone F isrepresented
although
there
isnofaulttraceat thesurface
in Figure
14=(seealsoFigure
3•).
TABLE 4. Fault Plane Solution Parameters Obtained From Inversion of P and $H

Waveforms
for the Kalamata(Figure9a) andthe Northwestern
Crete(FigureA1) Earthquake
Date Latitude Longitude Depth, Strike, Dip, Rake, Moment,

km deg deg deg N m
Apr. 27, 1965 $$ø42.0' 25ø$0.0

' 14.0 191.0 64.0 -79.0 2.0x1017
Sep.13, 1986 37006.0' 22ø11.4
' $.0 204.0 45.0 --77.0 ?.9X1017
to us that scenarii IIa and IIb are not very likely because to a magnitude4.5 or 5 event). Whether this eventis part
the absence of aftershocks around the epicentral area would of the main shock(i.e., occurredon the edgeof the fault
rather indicate a clean initiation of the rupture than a plane) or triggeredit (and occurredoutsidethe main fault
complex one. Thus the main shock may have been initiated plane,within the relayzone)cannotbe answered.In either
in the northerncluserby a smallenergyrelease(equivalent case, the main shock initiated within the zone offseting the
Kalamata and Thouria en dhelon normal faults which is
subjected to complex deformation. But once the rupture
front met with the main fault plane, slip could occur easily,
releasing most of the energy there. The observation of the
decreasing coseismicsurface displacement toward the south
fromsite 4 to site 1 (Figures2band4) is additionalevidence
Frictional
in favor of the southward propagation of the rupture. The
10-20km gentle way in which the surface rupture attenuated and
finaly stopped before the southernmost extremity of the
fault system(Figure 2) may be relatedto the low levelof
Quasi-plastic aftershockactivitythere(Figure
Deformation in the Rela•l Zone Between the Kalamata
and the Thouria Fault•
5.5:•M >•6.5 Focal mechanisms in the northern cluster show that in this
area deformation is complex, including focal mechanisms
with T axis azimuths oriented about Nl10 ø, N70ø, and
N20ø and with horizontal projection of slip vectors lying
at high anglesto that of the Kalamata fault (Figure 16).
Only a few of thesemechanisms (e.g., 89, 90, 108, and
111) couldbe interpretedas pure east-westright-lateral
strike-slip, with horizontal projection of slip vector parallel
to that of the Kalamata fault. Those mechanismsmay
correspond to motion on strike-slip segments between the
M•6.5 Kalamata and Thouria faults. Nevertheless,a largenumber
of observedmechanismsin this area cannot be explainedby
such rigid kinematics. They, however, are all consistent
with a uniform state of stress where •rt is vertical, •r3 is
horizontal and roughly perpendicular to the trace of the
._
. Kalamata fault, and •r2 is roughly parallel to it. Preexisting
planes of weaknessparallel to •r3 and dipping to the north
or to the south could then be reactivated as pure normal
faults. Such planes of weaknessmay exist within the relay
zone at depth, along an older zone of transversefractures,
Fig. 15. Relationship between earthquake magnitude, as discussed in the tectonic section.
rupture dimensions, and width of perturbed zone surroun-
ding a normal fault plane. The star indicates location of Implications for Regional Active Tectonics
hypocenter,the hatchedarea is the ruptured plane, the per- The pattern of Quaternary and active faulting in the
turbed zoneis showndotted.For smallearthquake(top) Aegean region has recently been compiled from various data
the width of the ruptured plane is smaller than the size of sources(i.e. geology,morphology,bathymetry,Landsat
the seismogenic layer(frictional)and the rupturedoesnot images, etc.) andsummarized by Armijoet al. [1986].The
reachthe surface.For largeearthquake(bottom)the ruptu- following discussioncombinesregional tectonic observation
red planereachesthe surfaceand its width is limited by the from Armijo et al. [1986]and new detailedobservations
size of the seimogeniclayer. Surfacebreakslength, amount presented in this paper(Figures1 and 17).
of slip, and width of perturbed zone increasewith magni- Our new observationsclearly indicate that the present
tude. For the Kalamata earthquakethe aspectratio of the styleof deformationin Kalamata(Figure2), and probably
ruptureis about1 (middle),andaftershocks areconcentra- in the entiresouthernPeloponnesus(Figure1), isdominated
ted in a very narrow zonesurroundingthe fault plane. by normal faulting and extensionorientedapproximatively
0 5 10km
I •
I I[a Hb
, i
Fig.16. Aftershockdistributionwithrespect
to faultgeometry.(Top}TheKalamataclusterisrelated
to the downwardextensionof the Kalamatafault. The northernclusterand the gap of aftershocks
are
withinthe relayzonebetween
the KalamataandThouriafaults.P (pluses)
and 3' (dots)axesfor all
focalmechanisms determined in thisstudyareplottedfor eachcluster.In the northerncluster,aftershock
mechanisms areseparated into threegroupsaccording to differences in 3'axesazimuth.The openarrows
indicatemeandirectionof extension;the solidarrowsindicatemeandirectionof compression in the
footwallcluster.(Bottom}Earthquake rupturepropagation. Sketches illustratethreepossible
scenarii
discussed in the text. The locationof the main shockis shownwith a star, within the northerncluster
for I, within the gap for IIa and IIb. The ruptured area is in gray in eachcase.
N90ø-l10øE. We have recalculated the focal mechanism southern Peloponnesusand Crete, the bathymetry indicates
and the depth of the April 37, 1965 earthquake(m•-$.7} that largeN-S normalfaultsexist[/•I/ber/•eta/., 1983]and
northwest of Crete using waveform modeling of long-period
that E-W extension may also be the dominant active regime.
P and $H waves(see the appendix). The earthquake The N-S normal faults, from southern Peloponnesus to
occurred within the crust at a depth of 14+$ km and Crete, representa rather uniform tectonicregime { •r3 -_-
correspondsto almost pure normal faulting along a N- E-W, •r3 -• N-S, •rz vertical) (Figure 1). They cut the
S striking plane. This plane may be associated with folds and nappe structure of Miocene age of the Hellenic belt
the normal faults that appear to control the large north- and thus are mostly newly formed faults. In places where
south bathymetric escarpmentson the westernedgeof Crete •r3 formsa relativelylargeangle(>45ø) with the Hellenic
(Figure 1}. Thus the combinedobservation of a normal structures, those structures may be largely reactivated by
fault earthquake mechanism and of active normal faulting quaternaryfaulting(Spartefault eastof the Taygetosrange,
indicates E-W extension there. Some of these faults are Figure 1). In contrast,where•r3 is almostparallelto the
also observed on land, particularly at the edges of the two Hellenic structures, the normal faults cut at high angle the
N-S peninsulasin westernCrete (Figure 1). Between old structuresand are thus independentof them (Crete).
onnesus
cm/yr
200km
M
24 ø 32 ø
a
b
Fig. 17. (a) Sketchmapof fault kinematicsin the Hellenicarc. Faultsare simplifiedfrom Figure1 and
Armijo et al. [1986].Smallsolidarrowsshowdirectionof localextension.En •chelonN-S normalfaulting
imply right-lateral motion between Peloponnesus and western Crete and left-lateral motion between
Anatolia and easternCrete (openarrows).Averagedirectionof activeextension(•3) is about N-S at
and north of the volcanic arc and about E-W in the Hellenic arc, near the subduction zone. The arrow
M indicatesthe averageazimuthof slip vectorof subduction earthquakes (N31øE) [McKenzie,1978].
The relativeAfrica/Europeconvergence (_• 1 cm/yr) is from Chase[1978].(b) Outlineof the evolution
of normal faulting in the Aegeanregionfrom late MJocene-Pliocene (?) (left) to Pliocene-Quaternary
(right). Largeopenarrowsrepresentthe bulk Aegeanextension.NAF is the North Anatolianfault, A is
the collisional boundary with Apulia, other symbols as in Figure 17a.
Becausethe Hellenicstructurefollowsthe trench(andforms 1977]or to forcesactingat the baseof the lithosphere

[e.g.
the Hellenicarc), any intermediatesituationbetweenthese McKenzie,1978]. In Crete, normalfaults parallel to the
two extreme cases may be found. arc are observed but microtectonic measurementssuggest
The observations that the E-W active extension is of that the direction of extension changed from N-S to E-W
regionalimportance(Figure1) callsfor a discussion
of its [Angel•er,1979]. Our resultsimply that the extension
relationship with the active N-S extensionoccurring in most perpendicular to the Hellenic arc is no longer dominant
of the Aegean. This N-S extension, which can be observed and that older (Late Miocene-Pliocene?)
normalfaulting
now in the Gulf of Corinth only about 100 km from the parallel to the arc probably formed prior to the onset of
Taygetosfaults (Figure 1), is roughlyperpendicularto the recent(Pliocene-Quaternary?)
N-S normalfaultingwhich
Hellenic arc and to the active volcanic arc and has often overprints it.
beenlinkedto large-scale
boundaryforces[e.g. Tapponnier, The geometry of N-S normal faulting is characterizedby
large•chelons. A clearexampleis the fault systemalong This change does not require a large modification of the
the westernedgeof the Taygetos.
(and easternedgeof the regionaltectonicstress(Figure 17). The whole Aegean
/Figure1) whereat leastthreeandproba- regionbeingdominatedby extension(•rl vertical),slight
gulfof Messenia)
bly fourmainsegments (Thouria,Kalamata,andKardha- differencesbetween the values of horizontal principal stresses
mill/Figure2a)andpossibly
another (Figure may
oneoffshore lead to different faulting patterns. Probably, the
1)) canbe distinguished.Thosesegments probablycor- amount of E-W extensionnear the subductionzone (_•
respondto a morecontinuous, mostlyreactivated structure 10%?)is smallcomparedto the bulkof N-S extensionin the
at depth. The kinematics of surfacefaultingrequiresthe rest of the Aegeanregion(•_ 100%) [e.g. McKenzie,1978;
averagedisplacement on this structureto be oblique,with Le P•chonandAngel•er,1979]. Nevertheless, the pattern of
a right-lateralcomponent of slip. The smallright-lateral active N-S normal faulting may be viewed as the responseof
component of slip observedin slickensidesalongthe Kala- the overriding plate to the horizontal force transmitted by
mata fault is consistentwith this average oblique displace- the incoming African margin acrossthe Hellenic subduction
ment(Figure5d). Rotationof blocks
withinthelargeright- zone. Then the average strike of the newly formed normal
lateral en •chelonsystemwouldrequire,however,smallcom- faults (•_ N10øE) is an estimateof the directionof this
of left-lateralmotiononthe individualnormalfault force (along•r2 in Figure 17a). The averageazimuth
ponents
segments
[e.g.,Jacl•son
andMcKenzie,
1988]. The left- of slip vectors of thrust earthquakes along the entire arc
lateral componentimplied by the main shockmechanismof (N31øE [McKenzie,1978;Jacl•son andMcKenzie,1988],M
the Kalamata earthquake would thus be in better agreement in Figure17a) is slightlydeviatedfrom this directionbut is
with such rotational movements. based on a limited number of tightly constrained fault plane
At a larger scale, betweensouthernPeloponnesus and solutions. The average direction of Africa-stable Europe
convergence
Crete,the overallpatternof faultingis alsoen •chelonand (_• N6øE [Chase,1978])is roughlyparallelto
oblique
to theHellenicsubduction 1 and17a). the average strike of normal faulting near the subduction
zone(Figures
This alsorequiresan overallright-lateralcomponentof slip boundary. This suggests a quite simple pattern for the
to be absorbed within a wide zone parallel to the trench late Neogenekinematicsof the Aegeanregion(Figure 17b)
(Figure17a). The observations in this paper where lateral movements, originated at distant collisional
discussed
suggestthat this en •chelonstructureis formedof almost boundaries, are apparently balanced with, on one side,
purelynormalfaults whichreflectE-W stretchingoblique the westward displacement of Anatolia, and on the other
to the arc. That the Hellenic belt is segmentedby those side, the east or southeastward push exerted by Apulia
faults and that a large part of it is now below sea level [Tapponn•er,
1977]. The forcetransmittedby the African
betweenPeloponnesusand Crete corroboratethis inference margin impeeds further N-S stretching and creates a new
andrequirea significantamountof E-W extensionacrossthe fault pattern only near the subduction zone and not in
et al., 1982],probablyof the orderof several the rest of the Aegean.
belt [Lllber•s This corroborates the earlier
kilometers across an E-W section of the en •chelon zone. inference that the Aegean lithosphere is particularly weak
On the eastern side of the Hellenic arc, between eastern [e.g., Tapponn•e•,1977; McKenzie, 1978; Le P•chon and
Crete and Anatolia, Quaternary faults observedon satellite Angel•e•,1979].
imagesand inferredfrom bathymetricchartsas well as mi-
crotectonic
measurements [An•lel•er,
1979]suggestthe exis- APPENDIX: DEPTH AND FOCAL MECHANISM
tence of another large scale•chelon array of approximately OF THE APRIL 27, 1965, EVENT NORTHWEST
N-S strikingnormalfaults (Figure17a). This structure OF CRETE(35.7øN,23.5øE,rob=5.7)
wouldrequire,ason the westernsideof the arc, a significant
amount of E-W extension across the mountain belt and a McKenzie[1972]publishedfor this eventa focalsolution
"conjugate
• left-lateral componentof slip within a zonepa- computed by placing the event in the mantle at a depth
rallel to the trench. All these observationssuggestthat both of 50 km and basedon first motionreadings(strike=214
ø,
the eastern and western •chelons are not formed by strike- dip=50ø, rake=-80ø). Herewe presentthe resultsof the
slip fault segments associatedwith the southward displace- inversion of the P and $H available WWSSN long-period
ment of Crete with respect to Peloponnesus and Anatolia, waveforms of this earthquake. These data clearly show
asproposed by Angelieret al. [1982]. that this event is shallowand its mechanism,althoughnot
The changein the tectonicregime(from an extension very well constrained, differs somewhat from that proposed
perpendicular to the arc, mostly N-S, to an E-W extension by McKenzie[1972](FigureA1). Oneplaneis moderately
acrossthe belt) suggests a changeof unprecised Pliocene well constrained(strike= 1914- 15ø, dip= 65 4- 10ø, Figure
age in the boundary conditions. Because this change is A2), and $H waveformsat CHG and QUE in particular
observed only in the southern Aegean area, close to the are usefulin constraining the valueof the rake(-79 q-20ø,
subductionzone(lessthan 200 km, Figure17a), boundary Figure A2). Using a crustal P velocityof 6.2 km/s, the
conditions have most probably changedthere. This could be depth is found to be 134.5 km. The crust in the area
one effect of the incipient collision of the Hellenic arc with is a thinned continental crust and its thickness, based on
the African margin to the south. North-south shortening inrefractionprofiles[Makris,1978],is believedto be about 30
the Mediterraneanridge[e.g.,Rllanet al., 1982]is probably km in Crete and 20 km farther north in the sea. Although
an other effect of such a collision. The increasingly buoyant the refraction profiles run north and east of the area of
crust of the African margin being subducted would hold the interest here, the crust is probably not much thinner and
southern part of the arc. Then, the Hellenic arc and the thus the earthquake occurred within the crust. The focal
adjacent Aegean crust would have to spread toward places mechanism found here is in agreement with the polarity
where less buoyant crust is being consumed as, for instance, readingspresentedby McKenzie[1972];however,the two
the Ionian basin to the west. nodal planeshave been rotated about 30ø to the north.
14,998 LYON- CAEN ET AL.: THE 1986• KALAMATA• EARTHQUAKE
CRETE 1965.04.27
NA SB
L L
CC CE
H CC
M
(3 ".,•
QD
5s STF
5'0s
QE
U
E
,•
pF
NH AG NB
A
I •,• • A
Fig. A1. Focal mechanismof the April 27, 1965 Crete earthquake obtained by simultaneous
inversion
of/• and$//WWSSN long-period
waveforms.
Observed
(solidcurve)andsynthetic(dashed
curve)seismograms
areplotted,scaled
to the seismic
moment.Earthquake
parameters
usedfor
the computation of synthetic seismogramsare listed in Table 4. STF is the source time function.
The amplitudescaleis in microns.Open squareis the 3' axis, solidsquareis aø axis.
NILP NAIP CHGSH OUESH NAISH CMCSH
Fig. A2. Sensitivity of waveformsto variations in strike and rake at six selectedstations. Row 1
is the solutionshownon FigureA1. Rows2-5:oneparameter wasfixed(rows2-3:strike,rows4-5:
rake)andwaveforms wereinvertedfor the remaining
ones.Numberson topof the focalspheres
are
strike/dip/rake/depth/moment
x1016.
LYON- CAENET AL.: THE 1986,KALAMATA,EARTHQUAKE 14,999
Acknoa/edipn•nt,.
We thank the EarthquakePlanningand Pro- empirical relations in seismology,
Bull. $dsmol.$oc.Arno,t•5,
tectionOrganization
(OAP.H) of Greeceandthe GreekArmy for 1073-1095, 1975.
their efficienthelp during field work and P. Bernard, A. Des- King, G., Z. Ouyang,P. Papadimitriou,A. Deschamps,
J. Gagne-
champs,and D. Hatzfeldfor fruitful discussions.R. Madariaga pain, G. Houseman,J. Jackson,C. Soufieris,and J. Virieux,
and P. Tapponniercriticallyread the manuscriptand suggested The evolutionof the gulf of Corinth(Greece):An aftershock
improvements.Drafting was done by G. Aveline. This work was study of the 1981 earthquakes, Geophys.J. R. Aetron• $o½., 80,
funded by the Institut National des Sciencesde l'Univers and we 677-693, 1985.
thank G. Aubert and J.C. Rossignol
for its promptorganization. Lee, W.H.K., and J.C. Lahr, HYPO71: A computer program for
This is contribution IPGP Nø1041.
determining hypocenter, magnitude and first motion pattern
of local earthquakes, U.S.Geol.Surv. Open File Rep.,pp. 75-311,
REFERENCES 1975.
Angelier, J., N•otectonique de l'arc •g•en, th•se d'etat, 405 pp., Le Pichon, X., and J. Angelier, The Hellenic arc and trench
Univ. Paris VI, 1979. system: A key to the neotectonic evolution of the eastern
Angelier, J., N. Lyberis, X. Le Pichon, E. Barrier, and P. Huchon, Mediterranean area, Tectonophysies, 60, 1-42, 1979.
The tectonic development of the hellenic arc and the sea of Lyberis, N., Gdodynamique du domaine •gien depuis le Miocene
Crete: A synthesis, Te•tonophyzi½%
86 159-196, 1982. supirieur, th•se d'itat, Univ. Paris VI, 1984.
Arabasz, W.J., R.B. Smith, and W.D. Richins, Earthquake stu- Lyberis, N., J. Angelier, E. Barrier, and S. Lallemant, Active
dies along the Wasatch front, Utah: Network monitoring, seis- deformation of a segment of arc: The strait of Kythira, Hellenic
micity and seismic hazards, Bull. Seisrnol.Soc.Am., 70, 1479- arc, Greece, J. Struct. Geol., 4, 299-311, 1982.
1499, 1980. Makris, J., The crust and upper mantle of the Aegean region from
Armijo, R., A. Deschamps, and J.P. Poirier, Carte sismotecto- deep seismic sounding, Tectonophyzics, 46, 269-284, 1978.
nique de l'Europe et Bassin m•diterran•en, Inst. de Phys. du McCaffrey, R., and J. N&b•lek, Earthquakes, gravity and the
Globe, Paris, 1986. origin of the Bali Basin: An exemple of a nascent continental
Bernard, P., and A. Zollo, The Irpinia, Italy, 1980 earthquake: fold and thrust belt, J. Geophys.Res., 9•, 441-460, 1987.
Detailed analysisof a complexnormal faulting, J. Geophys.Res., McKenzie, D., Active tectonics of the Mediterranean region,
in press, 1988. Geophys.J. R. Actton. Soe., 30, 109-185, 1972.
Bounif, A., H. Haessler, and M. Meghraoui, The Constantine McKenzie, D., Active tectonics of the Alpine-Himalayan belt: the
(northeastAlgeria) earthquakeof October, 27 1985: Surface Aegean Sea and surrounding regions, Ge•phys.J. R. Actton.
ruptures and aftershocksstudy, Earth Planet. SO/.Left., 85, 451- 55, 217-254, 1978.
46O, 1987. Meissner, R., and J. Strehlau, Limits of stresses in continental
Chase, C. G., Plate kinematics: The Americas, East Africa and crust and their relation to the depth-frequency distribution of
the rest of the world, Earth Planet.So/.Lett., $7, 355-368, 1978. shallow earthquakes, Tectonics,1, 73-89, 1982.
Chen, W.P., and P. Molnar, Focal depths of intracontinental and Mercier, J.L., N. Delibassis,A. Gauthier, J.J. Jarrige, F. Lemeille,
intraplate earthquakes and their implications for the thermal H. Philip, M. Sibrier, and M. Sorel, La ndotectonique de l'arc
and mechanicalproperties of the lithosphere,J. Ge•phys.Res., igten, Rev. Geol.Dyr• Gdogr.Phys., •1, 67-92, 1979a.
8œ, 2945-2969, 1983.
Mercier, J.L., N. Mouryaris, C. Simeakis, T. Roundoyannis, and
Delibassis,N., J. Makris, and J. Drakopoulos,Crustal and upper C. Angelidhis, Intraplate deformation: A quantitative study
mantlestructureof westernGreece,Ann.Geol.PaysHell.,33 of the faults activated by the 1978 Thessaloniki earthquakes,
47-61, 1988. Nature, œ78,45-48, 1979b.
Deschamps,A., and G.C. King, Aftershocksof the Campania- N&b•lek, J., Determination of earthquake sourceparameters from
Luciana(Italy) earthquakeof 23 November1980,Bull.Se/smol. inversion of body waves, Ph.D. thesis, 361 pp., Mass. Inst. of
So½.Arr•, 74, 2483-2516, 1984.
Technol., Cambridge, 1984.
Deschamps,A., G. Iannaccone, and R. Scarps, The Umbrian
Ouyed, M., G. Yielding, D. Hatzfeld, and G.C. King, An after-
earthquake
(Italy) of 19 September
1979,Ann.Geophys., shockstudyof the El Asnam(Algeria) earthquakeof 1980Oc-
29-36, 1984.
tober 10, Geophys.J. R. Actton. Soe., 73, 605-639, 1983.
Dziewonski, A.M., and D.L. Anderson, Preliminary reference
Papazachos,B., A. Kiratzi, B. Karacostas, D. Panagiotopoulos,
Eart•hmodel,Phys.EarthPlanet.Inter.,•5, 297-356, 1981. E. Scordilis, and D.M. Mountrakis, Surface fault traces, fault
Dziewonski,A.M., and J.H. Woodhouse, An experimentin syste- plane solution and spatial distribution of the aftershocks of the
matic study of global seismicity:Centroid-moment tensor solu-
tions for 201 moderate and large earthquakesof 1981, J. G•o- September13, 1986earthquakeof Kalamata(southernGreece),
phys.Res., 88, 3247-3271, 1983. Pure Appl. Ge•phys.,I•6, 55-68, 1988.
Dziewonski, A.M., G. EkstrSm, J.E. Franzen, and J.H. Wood- Richins, W.D., J.C. Pechmann, R.B. Smith, C.J. Langer, S.K.
house, Centroid moment tensor solutionsfor July-September Goter, J.E. Zollweg, and J.J. King: The 1983 Borah Peak,
1986, Phys.Earth Planet.Inter., 46, 305-315, 1987. Idaho, earthquake and its aftershocks,Bull. Sdsmol.Soc.Arr•,
77, 694-723, 1987.
Granter, T., Origin, damping, and pattern of developmentof
faults in granite, Tectord½%4, 721-738, 1985. Ryan, W.B.F., K.A. Kastensand M.B. Cita, Geologicalevidence
concerning compressional tectonics in the eastern Mediterra-
Hatzfeld, D., A. Christodoulou,E. Scordilis,D. Panagiotopou-
los, and P. Hatzidimitriou, A microearthquakestudy of the nean, Tectorwphyzics,86, 213-242, 1982.
Mygdoniangraben(northernGreece),EarthPlanet.So/.Left., Scholz,C.H., Scalinglawsfor largeearthquakes.Consequence on
81, 379-396, 1987. f physicalmodels,Bull. Setstool.Soc.Arno,7•, 1-14, 1982.
Instituteof GeologyandMineralExploration,Geological mapof S•brier, M., Tectoniquer•cente d'une transversale• l'arc •g•en,
Greece,Kalamata sheet,scale1:50000, Athens,1986. (le golfede Corintheet sesr•gionsp•riph•riques), th•sede
Jackson,J.A., and D. McKenzie, The geometricalevolutionof 3•me cycle, Univ. d'Orsay, France, 1977.
normal fault systems,J. S•. Geol.,5, 471-482,1983. Soufieris,C., and G.S. Stewart, A sourcestudy of the Thessaloniki
Jackson, J.A., andD. McKenzie,The relationship betweenplate (N. Greece)1978 earthquakesequence,G•ophys. J. R. Actton.
motions and seismic tensors, and the rate of active deformation Soc., 67, 343-358, 1981.
in the Mediterranean
andMiddleEast, Geophys.
J., 93,45-73, Soufieris,C., J.A. Jackson,G.C.P. King, C.P. Spencer,and C.H.
1988.
Scholz,The 1978earthquake sequencenearThessaloniki (nor-
Jackson,J.A., J. Gagnepain,G. Houseman,G.C. King, P. Pa- thern Greece),Ge•phys.J. R. Aetron•
So½.,
68, 429-458,1982.
padimitriou,C. Soufieris,and J. Virieux, Seismicity,normal Tapponnier,P., Evolutiontectoniquedu systemealpin en Midi-
faultingand the geomorphological development of the gulf of terran•e: Poinqonnementet •crasementrigide-plastique,Bull.
Corinth(Greece):The Corinthearthquakes of Februaryand So½.gdol.•-k, 7, 437-460, 1977.
March 1981,EarthPlaNet.$ci.Lett.,57, 377-397,1982. Tchalenko, J.S., Similarities between shear zones of different
JaegerJ.C., andN.G. Cook,F•ndarne•tals o!RockMechavdcs,593 magnitudes,Geol.So½.Arr• Bull., 81, 1625-1640,1970.
pp., Chapman and Hall, London, 1979. UNESCO,InternationalB•thymetryChart ofthe Mediterranean,
Kanamori, H., and D.L. Anderson, Theoretical basis for some scale 1:1,000,000, 1981.
Westaway, R.W.C., and J. Jackson, The earthquake of 1980 N. Delibassis, V. Kouskouna and K. Makropoulos, Division of
November 23 in Campania-Basilica, Geophi/•.Y. R. A•tron. $o½., Geothermy, University of Athens, 15784 Ilissia, Athens, Greece.
9/), 375-443, 1987. R. Gaulon, H. Lyon-Caen, and P. Papadimitriou, Laboratoire
Zandt, G., and T.J. Owens, Crustal flexure associated with nor- de Sismologie, Institut de Physique du Globe de Paris, 4 place
mal faulting and implications for seismicity along the Wasatch Jussieu, 75252 Paris Cedex 05, France.
fault, Utah, Bu/l. $e/srno/.$oc.Ant, 70, 1501-1520, 1980. G. Pedotti, Observatoire de Grenoble, IRIGM, B.P. 53X, 38041
Grenoble Cedex, France.
R. Armijo, Laboratoire de Tectonique, Institut de Physique du
Globe de Paris, 4 Place Jussieu, 75252 Paris Cedex 05, France.
J. Baskoutass,J. Drakopoulos,J. Latoussakis,and D. Papanas- (ReceivedDecember4, 1987;
tassiou, SeismologicalInstitute, National Observatory of Athens, revised July 26, 1988;
Athens, Greece. acceptedAugust 17, 1988.)
LYON- CAEN ET AL.: THE 1986,KALAMATA, EARTHQUAKE 15,349

The 1986 Kalamata (South Peloponnesus) Earthquake: Detailed Study of A Normal Fault, Evidences For East-West Extension in The Hellenic Arc

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The 1986 Kalamata (South Peloponnesus) Earthquake: Detailed Study of A Normal Fault, Evidences For East-West Extension in The Hellenic Arc

Uploaded by

Copyright:

Available Formats

The 1986 Kalamata (South Peloponnesus) Earthquake:

Detailed Study of a Normal Fault, Evidences for

To cite this version:

HAL Id: hal-01994159

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

H. LYON-CAEN,1 R. ARMIJO,2 J. DRAKOPOULOS,3'4 J. BASKOUTASS, 4

INTRODUCTION Because of the low level of seismicity compared with that

1982;Lyberis,1984]. In contrastwith this apparentunifor- late Helleniccompression

:'*:''[..•:i• '-'•' '• ?.•ry"-•.•....

::• ............ ......

,• ....•f':• ..•'"'; ' ?•:"------:-:-::

5 k.m "i '''.................................

The coseismicmovement correspondssystematicallyto

Fig. 2a. Map of activefaultsnearKalamata.Topography(in meters)and geologyare simplifiedfrom

37o00 •' 37o00 '

offset between the Kalamata and Thouria faults is located

Locationsof the main shock(September13, 1986,at 1724

Fig. 5. Detailsof thefaultbreaksouthof Eleochori

piedmont sediments Cretaceous-Eocene

Fig. 6. Schematic crosssectionof the Kalamatafault nearsite 2 (Figure2b). Circleblowsup the

velocitystructurenot taken into accountby the preliminary pendentof the assumed

TABLE 1. Hypocentral Parameters

Latitude Longitude Depth, km ML Ms rnb

Main ShockSeptember15, 1986, 1724 GMT

Athens(NOA) 37ø 6.0' 22011.4 ' 1.0 5.5 6.4

AftershockSeptember15, 198t•, 1141 GMT

Athens(NOA) 37ø 4.8' 22004.2 ' 1.0 4.8 5.3

tion which disagreeswith observations. Moreover, polarities

37•o 1POL ,0 , 2 , 4Kin--

37 3 -c/-. •_• • ¾j,• •. -'v _ 37 3

Fig. 8e. Map viewof the 344 bestlocatedaftershocks

KALAMATA 1986.09.1,5 to study the influenceof the Vp/V• ratio we foundthat it

KEY P HYB P SLR P CHO SH BUL SH BEC SH KBS SH

Focal Mechanisms TABLE 2. Station Coordinates and Velocity Structure

andcrosssections perpendicular(N105ø) to the meanstrike

5!0 5!2 5!4 5!6 5!8 6!0 Vp(km/s)

2.8 2.7 2.8 2.9 :30

Fig. 11. Focalmechanisms of 133 aftershocksplottedon the lowerhemisphere. Solidcirclesare

Fig. 11. (continued)

101 102 10S 104 105

108 107 108 10• 110

126 !27 128 129 1SO

Fig. 11. (continued)

0 KKAL (•/• • 11o

Fig. 12. Map views of 133 focal mechanisms

Fig. 12. (continued)

Fig. 12. (continued)

94, 97, 99, 100, 101, and 103), the othermechanisms

Fig. 12. (continued)

N15=E to the presence of a large number of aftershocks and to a

Kalamata Fault Kalamata Fault

TABLE 4. Fault Plane Solution Parameters Obtained From Inversion of P and $H

Date Latitude Longitude Depth, Strike, Dip, Rake, Moment,

Apr. 27, 1965 $$ø42.0' 25ø$0.0

Becausethe Hellenicstructurefollowsthe trench(andforms 1977]or to forcesactingat the baseof the lithosphere

NILP NAIP CHGSH OUESH NAISH CMCSH

You might also like