Professional Documents
Culture Documents
Geodynamics of Flat Subduction: Seismicity and Tomographic Constraints From The Andean Margin Guts Cher 2000
Geodynamics of Flat Subduction: Seismicity and Tomographic Constraints From The Andean Margin Guts Cher 2000
E. Robert Engdahl
Department of Physics, University of Colorado, Boulder
814
GUTSCHER ET AL- GEODYNAMICS OF FLAT SUBDUCTION 815
Carib. Plate
10øN
Cocos
•• - ...
Plate
Eastern
Cordillera
South
American o
Inca Plate
7.0
Plateau
Brazilian
10øS
Shield
7.Bcrn•===•
Bolivian
Subandes
Nazca
20øS
Plate
Sierras
Pamp-
8.4 eauas
30øS
Juan
Fernandez
R.
:....
Inca
Plateau
140 km -J
/
Figure 2. Three-dimensional
view (lookingsouth)of the subductedNazca slab surfaceas
determined from gridding of earthquake hypocenters. The location of oceanic plateaus at the
surfaceand their estimatedsubductedpositions(dashed)are shown.Two morphologic
highscan
be seenwith an intervening "sag". The highscorrespondto the subductedcontinuation of Nazca
Ridge and to the subductedInca Plateau [Gutschefet al., 1999a].
[km]
0 100 200 300 400 500 600 700 800 900 1000
,. , , Sec,on:S.cuor-moer,e,.,ee,.,,vo,
c.rc
• -8ø
-t!' ''*•;:ii:"•i•!::•:;•;.:"•
-œ;)"ø
'møderatel
' • •'•"
"-•:--•:'•••••i
' ii"•i:iiiii•'
'"ii;i:'i
• ,. -so
I •, q::::::..
"•'-:"%'k:...............
-';;?':•-.-':"•'-
:'::"'"• :LL:_•.•
-•;ø':•'e;-.•'
-':'•
•..... z•oo
'W-----'•"•'•:'•'•••.--••••:.••::.:....:.....:,:':•.i•:!
'- '--
lo
• -12o
• ,•_ .......... ..-••:::•::::•:;.;•j•.::%;
-• -.•-..•.• .......
• -160
e •,•
-2•
"sag"
region
,,"
Section C: Central Peru - interm. subduction, "sag"
!...iii:iii:•i'?,'i•!?"'
..::..i!?iiii•iii!'i;i•½•i'•'"":'•?
•.....................
'½':"•½•i
-4of"'""•:
:':•'•:'"•'•":':•i•-•::':.;'•E':
-..• % o•o,-•;, ,
I , •--•::• .v'• •.. ,
-8o
J_
-120
460
___
' ½'%•'•':'•?.
Nazca
flat:
slab
Ridge"•
-2• - I
o
NazcaRid'""•"•••:•••••'-'""•-..-."•
:•..•..
'.'.••:••.•..••.:.:..•••
b'
-8o , .....
:.•;;•fz•,:•::•;•i•E.•.•:•:•?•::½•;:::•:j•:::;::
__
-•2o
-160
-2•
ß o 10
-2• 0 5
• •o
....... '......
• , ø'::?•'ed%•ff&%:.ø
steep
slab •-'==============
• -16o ""-•• T
Figure 3.
818 GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION
A series of 10 ENE oriented 2-D cross sections per- densermaterial. Earthquakes are shownas small white
pendicular to the trench and to the Andes are presented circles.
in Figure3 (left), eachwith a samplingwidth of 1ø lat- The verticalcrosssections(Figure3, right)constrain
itude. They cover, from north to south, a moderately the geometryof the downgoingNazca Plate in threedi-
steepslabin southernEcuador(section1), the flat slab mensions.They correspondto seismictysections1, 3,
due to Inca Plateau (sections2 and 3), the sagregion 6, 8 and10respectively.
In southern
Ecuador(Figure3,
(sections4-6), the flat slabdueto NazcaRidge(sections sectionA) the NazcaPlate dipsmoderatelysteeplyto-
7 and 8) and the steepslab in southernPeru/Bolivia wardthe NE, curvinggentlydownto a depthof 600km,
(sections9 and 10). While the sagis not a major first- where the great 1970 Colombiaearthquakeoccurred.In
order feature, earthquakesin this region are roughly 20 northern Peru the flat slab near Inca Plateau is sam-
km deeperthan thosein the flat slabto the north (Inca pled and is shown to "underslide" the South Ameri-
Plateau) and 40 km deeperthan in the flat slabto the canlithosphere
horizontally
for roughly800km (Figure
south (Nazca Ridge). The maximumdistanceof pene- 3, sectionB). The lack of continuitybetweenthis flat
tration of the horizontal slab correspondsto the north- slab portion and the subductedlithosphereat 500-600
ern Inca plateau. Here the horizontal plate bendsdown km depth suggestsslab detachmenthas likely occurred
steeply or "resubducts" some 700 km from the trench here, perhaps along an age discontinuityas described
(section2), whereasin the sagregion(section6) resub- by previousauthors[Vlaar and Wortel,1976; Wortel,
duction occurs 500-600 km from the trench. Although 1984].Spiketests(not shown)indicatemoderatereso-
the continuation of Nazca Ridge representsthe shallow-lutionat thesedepths(200-400km) andsuggest theslab
est portion of the flat slab, at depths between 80 anddetachment imagedis not an artifact.In the sagregion
100 km (sections8 and 9), it is not possibleto deter- (Figure3, sectionC) the overallplatedip is transitional
mine the distance of penetration basedon hypocenters between the flat slab north and south. Here the Nazca
alone, owing to the seismicgap here, though the to- Plate curvesgently,becomingsteeperasit descends
and
mographicimage(Figure3, sectionD, seealsobelow) is relativelycontinuous.Near NazcaRidge(Figure3,
suggestspenetrationof up to 800 km. sectionD) the slab is again shownto proceedhorizon-
tally, underslidingthe SouthAmericanlithosphereover
2.2. Tomographic Model 500km beforebendingsharplyat a nearly90ø angleand
Tomographicmodels can reveal information about descendingalmost vertically to the 670 km discontinu-
lithosphericand mantle structure at great depths and ity. Finally, the section beneath southern Peru reveals a
in regionswhereno informationis availablefrom in situ steepslab segment,descendingcontinuouslyand curv-
seismicity[En9dahlet al., 1•)95].Improvements
in the ing smoothlytowardthe 670 km discontinuity
(Figure
inversion techniquesof global earthquake travel time 3, sectionE).
data now allow resolution comparable to regional to- The horizontalsections(Figures4a-4d)revealveloc-
mographicstudies[Bijwaardet al., 1•)•)8].The tomo- ity variations in layers of 50 km thickness. Wadati-
graphicvelocitymodel is a globalmodelobtainedby a Benioff contoursare also shownbasedon an analysis
travel time inversion based on 82,000 well-constrained of hypocenters[Gutschefet al., 1999a]and correlate
earthquakesfrom a global relocated hypocenterdata well with the depth range of the high-velocityanoma-
set [Engdahlet al., 1•)•)8],using7.6 million P and pP lies. The abrupt northern termination of the flat ly-
arrivals with absolute residuals <7.5 s. The descend- inglithospheric
slab(topleft cornerof Figure4b) sug-
ing oceaniclithospherecan be seenas zonesof higher gestsslower,warmer material to the NW and supports
seismicvelocities(darkestshading)due to the colder, the hypothesis of a lithospheric tear beneath south-
.....;}:,
'•'•';'":••:: • ".... ....
•:: •.:
.-.'•,...:•-.:?
....... %V:-': •. • -
.: :--•'..•-.•
•
':•...... , , .Inca Plateau. -:•:.....•
...... :•--•
..... :' ;Inca Plateau
:;,:.::;:
]- -..•.'.:.....,. .,.:• - i • ::•'
;................
:.......,..a**..:..'*<'•>,;..,.,,
.... .:.....
:•;,..
,....,¾
::...,
......,,.
::....;.•
............
:_..$&:...:.:.,:2..,aL:
•-"• ........
..**;**., : :•';::..
'•'?•' **..a::, :.:.;:•._..L•......•-*•-•.•.•'
-:a;•,•,:.:.**:.:,:-----•:•;,•
.... '•;•;•:•*'
'•' / ' • • ;'":
....•:/•&...-.•;.;$
.............
•:.
,:.•:::,:;**•:
-;-•:• .:..-,.
....... . .-. ....**.--•,....s .
...............
..•.•..... .•,;:::
:":•.:•
.... '/• .,•,? .. '..-.• :..•
.....-..•. . . ._.;;:•'•;........ .'--.-.•.
:.a ':*:*'..
':-*
.......ß
.... ':?':*-,:
"':": ;• ' :• ,:*--.,,
.':",' '*;•:.: *::::<
..... • ,.
}*•a"::•'•:•
, :'"'•;
.a • '.•:;':;•:' •
:'
.......
"•;----:".'-:'--•:•-"
....*'::••':.•'•
•- '*$•,....
......
•:' :.•a '•••••.;•:.::•'
"•'••••,::,•}' ' '•:.....%."
-•.:--..-'
•...',•.•.---::.-•
•......
,x':.,:..>"•;•;...
•................. '--•.
•-,'
•
'• .................................
".a2:a•::•2
.............
-................
-:::.::::4.
ß :•/•... ".• • • .......... . .:::;?.:..:.?
.... :.....'•'<..•..•.•: :•:;::a:c...
'•..•.•.x:.• •:.
ern Ecuador along the continuation of the ENE trend- comparativeanalysis.While the Peru flat slab segment
ing Grijalva fracture zone [Pennington,1981;Hall and is by far the longest,the central Chile fiat slab seg-
Wood,1985; Gutschefet al., 1999b]. ment extending from 28o to 32øS is equally well docu-
mented [Pilger, 1981; Sinalleyand Isacks,1987; Cahill
3. Discussion and Isacks, 1992; Kay and Abbruzzi,1996]. Here the
Juan FernandezRidge is currently subductingbelow
3.1. Flat Subduction: Other Regions
the margin at 31øS [Pilger, 1981; 1984; yon Huene et
Beforeaddressingthe question,What causesflat sub- al., 1997]. In contrastto Peru, only one buoyantstruc-
duction7,it is usefulto examineother regionswherethis ture is present, and accordingly,the fiat slab segment
phenomenonis observedand thus gain insightsthrough and volcanicgap are only 500 km in width. As in Peru
An. 6B
Colombia 7 ø N E. Cordillera Section A: NW Colombia - flat slab
0 100 200 300 400 500 600 700 800 900 1000 "••..•'"••••'••i .....•:•"",•i• .:=,
.........
-........
'............
0
•=
'-•=.
',.'•.;.,.'i!:.
-20 ?.=.-- •.•:= ,.•:.
-40
-60
-80
-100
-120
-160
-180
z• 0 •00 200 300 400 500 600 700 8o0 900 •000 "'"""'*-.--'-:•.$';;.?'-'""'"•'
.'.=• •'••.• •?'
".... ..--
0/ ' _ ' ___'.._ ' ..... '.....•-t -_----• k•
.......... • ß • • '•000 o "a:x"•*•'$
:=-'•
:•:•*:,<:•:'.'•o
4 ',::"•':•x'.,.'s
,,.. • ß' ßß
-60
-80t - '-(•?, An.
6b?'::'"
*•**•½•;•;'"'"•'•... [• Section
.:W . -steep
Colom,,aslab
-1O0 '*;**it•
-120 .,*';- '
-180 *'*•'
-2oo 15 -22?Ma crust
3,oO
,o,o
^.,0oh";o44;o
.,o
:o:.,oo
,,:o
9oo
lOOO '
60 ß-.•,:-:*:-:'.•
....... •.'• . - •. ':*•;';•--.:::•:
.:-•-.---.-.-.'.;-•'.S.
:;•::•'
-140 • .•
-160
-220
• •00 o *.................
o
Chile 31 o S Sierras
An.20 Pampeanas
4) , i , t , , • ,,• • • i ,
' -......
-40
ß •"*' SectionC:CentralClfile/ NWArgentina
- flatslab
-60
-8o ....
•':•G•'-'-":":":L...
%*"
% • • ', 24'•
-120
-140
-160
-18o
-2oo 43 -55Macrust ,
Chile41 oS 5 •0 •5
5) i , t , An.5B/5E
i , i , i , I , i , • •
01............
-60
-80
• :;.:%- ?:... ... :•.c:F:•:.•;*•.....:•:•:.•:•:..`•*:.;•.•::::•:::::;..:•.:•%?:•b.•.•:*•*•::•:•:•::•
.....
-160
-180
t 15-25
-200
-220 Ma
crust • , VE=
2'1 -1.5% •-'aJ•?•'""'"::'•::i')!•!:...
' +1.5%
Region
• L,bkm Thickened
Crust? Age,
• m.y. V,dcm/a Arce Adkf,m.y. Deform.
g
the Nazca Plate proceedssubhorizontallyfor 300 km at lithosphere[Suarez et al., 1990; Singh and Mortera,
a depthof about 100 km (Figure5). The tomographic 1991]. In southernAlaskathe accretionof the Yaku-
imagefrom centralChile confirmsthis well established tat Terrane, a buoyant block of thickenedcrust, along
flat slabgeometry(Figure5, sectionC). the easternmost Aleutian Trench causes a remarkable
There are several other subduction zones worldwide flattening of the angle of subductionas constrained
exhibiting an "abnormal"flat to low-anglesubduction by the great trench arc gap of 600 km, by abundant
geometry.In general,theseare not as well constrained earthquakehypocenters,and by wide-angleseismicdata
as central Chile or Peru and subsequently,are less well [Pcnnington,1984; Brocheret al., 1994]. The Juan
known. In nearly all these cases,as in Peru and Chile, de Fuca slab beneath Cascadia also descendsat a very
subductingplateausor arcsare associatedspatiallyand shallowangle[Crossonand Owens,1987],althoughin
temporallywith this anomalousstyleof subduction(Ta- this case, no thickened crustal block has been identi-
ble 1). In Ecuador,CarnegieRidgesubductsbeneath fied at present. At the Nankai trough, oceaniclitho-
the northernAndes[Pennington,1981; Gutschefet al., sphereflankedby the Izu-Bonin Arc and Palau-Kyushu
1999b](Figure1). In NW Colombia, thePanama/ChocoRidgeunderslides SW Japanat 50-70km depth [Hira-
Block is collidingwith South America [Pennington, hara, 1981; Sacks, 1983; Morris, 1995; Gutschefand
1981;McGear!let al., 1985]. A portionof the Caribbean Lallemand,1999; Gutschef,2000]. Finally, there is the
Plateau is subductingalong the North Colombia De- New Guinea thrust beneath Irian Jaya [Okal, 1999].
formed Zone toward the SE while others portions may This subduction system, with no known arc volcanism,
currentlybe underslidingColombiafrom the west [van also appearsto exhibit flat slab behavior,likely related
der Hilst andMann, 1994](Figure1). The tomographic to the subductingEuripik Ridge, a 2 km high oceanic
image(Figure5, sectionA) supportsthis flat slab in- plateau[McGearyet al., 1985].
terpretation for NW Colombia. In Costa Rica, Cocos In many cases of flat subduction an intermediate
Ridgesubductsat the terminationof the Middle Amer- depth seismicgap is present[McGcaryet al., 1985;Mc-
ica Trench [Gardner et al., 1992; Protti et al., 1994, Cann and Habermann, 1989], and thus the geometry
1995; Kolarsk!let al., 1995] (Figure 1). In southern of the slab below 60 km depth must be constrained
Mexicoa shortflat slab segment(200-300km downdip by indirect observation(e.g., tomographic,volcanolog-
length) appearsto bottom out at a shallowdepth of ical, etc.). In Colombiaand CostaRica a volcanicgap
only 50 km, implying an unusually thin continental correlates with the position of the subducting arc or
822 GUTSCHER ET AL.. GEODYNAMICS OF FLAT SUBDUCTION
o lOO 200 300 400 500 600 700 800 o lOO 200 300 400 5O0 6OO 7OO 8OO
o • [ s , , ,
lOO lOO
6os
Peru Cascadia 48øN
200 200
1 oo lOO
Ecuador 0ø Alaska 148øW
200 200
o
.....,.•s i [ , , , • ,
lOO lOO
20o
Colombia
7oN • 200
SW Japan 133øE
adjacent arc
o • • • ' ' ' i , , i i , i
lOO lOO
ridge, while in Ecuador a broad arc is present. In tion behavior.For Peru (Figure3, sectionsB and D),
Costa Rica/Panama, Ecuador, and SW Japan, anoma- centralChile (Figure5, sectionC), NW Columbia(Fig-
lous "adakitic" geochemicalsignaturesare foundin relic ure 5, sectionA) and southernAlaska(not shown)the
or present-dayarc magmas[Dcfant½tal., 1993;Monzicr tomographicimages confirm an underthrustingof the
½t al., 1997; Morris, 1995]. In central Chile, adakitic oceanic plate over several hundred kilometers at a low
signatures werefoundin the relic (6-12Ma) arc and at- angleas constrainedby the hypocenterdata. In regions
tributed to the shallowingof the subductingslab [Kay where relatively young, thus warm, oceaniclithosphere
and Abbruzzi,1996]. Suchadakitic magmashave been is subducting(Ecuador,CostaRica, Mexico,Cascadia,
attributed to partial melting of very young (<10 m.y. and SW Japan), no fiat slab is clearly visible. This
old) oceaniccrust [Derantand Drummond,1990;Pea- may be attributed to lack of resolution in the upper
cock½tal., 1994],althoughparadoxically,they are com- 100 km and/or the possibilitythat a relatively warm
monly found in many arcswhere crust of moderateage slab may not possesssufficientseismicvelocity contrast
(45-10m.y. old) is subducting(Table 1). This intrigu- with respect to the ambient mantle. In other regions
ing correlation between adakitic magmas and many fiat (e.g., westernNew Guinea)the spatialresolutionis too
slab segmentssuggestsa possiblelink to the evolution poor to draw inferences. In all cases,however, the to-
of fiat subduction,but this subject is beyondthe scope mographic model does not exclude the presenceof a
of thispaperandis discussed elsewhere [Gutschef½tal.,
fiat slab. On the basisof our analysisof (1) earthquake
•000]. hypocenterdata [Engdahl½tal., 1998], (2) the tomo-
The globaltomographic
model[Bijwaardet al., 1998] graphicimagesdiscussed above,and (3) publisheddeep
has been examined for the subduction zones listed in seismicdata, simplified sectionsof the subductingslab
Table 1, interpretedto exhibit fiat to low-anglesubduc- geometryhavebeen compiledfor all 10 of theseregions
GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION 823
:-:•:---':
3.•:-•:•--
•:-:::•:. ..•?-:::?•'•
•:•:t:
upper
mantle
33 km(3.30) ?•:iiii.i:il-;';i::i:;:'::::,.:::.:.:
::/:iii::":
ii.:i-:.
53km(3.30)
48 km (3.30)
:•.:•i:::•:.::
.::-....::-
..........................
-.................................
-. 68km(3.30) 5O
?as!henO•i•-•:.-
::':;';•;;;;;::;?::::}:,;;i::;i;
i::-;::::i::•::!::iiiii!::::iiii:::i::!::':::.::i::;::i•;::;;;::;::;:::i
'' . ................................
densities in [km]
[103kg/m3]
lOO
support itself for 25 m.y. old lithosphere,and for 50 ence(i.e., centraland southernChile). Published[Jor-
m.y. old lithosphere,the column is denserthan the dan et al. , 1983; $uarez et al. , 1983; Gutschef et al.,
surroundingasthenosphere and will thus subductnor- 1999b]and regionalearthquakefocalmechanisms (Fig-
mally. This appears to be the casefor the 10-15 km ure9) arepredominantly compressional(thrusttype)to
thick Hikurangi Plateau, embeddedin Cretaceousage transcurrent(strike-slip)with P axestypically aligned
lithosphereand thereforesubductingbeneathNorth Is- parallel to the plate motion (east-west)suggestingthe
land New Zealand at a steep angle [Davy and Wood, subduction plate boundary stressesare transmitted far
1994;Ebevhavt-Phillips andReyncvs,1997].Conversely, (severalhundredkm) into the heart of the upperplate.
thicker plateaus will presumablybe able to support We attribute this increased seismicity to an overall
even wider fiat slab segments. The above valuescor- increasein viscousinterplate coupling betweenthe two
lithospheres(Figure 9). Becausethe interplate inter-
respond to the range of crustal thicknessesreported
for severaloceanicplateausworldwide,Galapagos(14-
faceis up to 400-500km longfor fiat subduction(com-
18 km)[FeighncvandRichards,1994],Tuamotu(30-32 paredto 100-200km for steepsubduction),platebound-
km) [Talandierand Okal,1987],CocosRidge(18 km) ary forcesare transmitted more effectivelyto the upper
[Bialaset al., 1999],Nazca Ridge (17 km)[Bialas et plate. The fact that the increasein seismicenergy re-
al., 2000],Marquesas(14-18km)[Caresset al., 1995], leased(a factor of 3-5 times greater for fiat slab seg-
and Kerguelen(20-30km)[C harris et al., 1995],sug- ments) correspondswell to the increasein interplate
gestingthat these globally abundant structuresare ca- contact area suggestsa causal link between these quan-
pableof alteringthe modeof subductionof moderately titires as proposedpreviouslyby other workers[Cross
old <60 m.y. crust. Most of the fiat slab segments and Pilger, 1982]. While it is generallybelievedthat the
aroundthe world causedby singleplateaus,(e.g., cen- base of the lithosphere deforms plastically and therefore
tral Chile, Ecuadoror CostaRica) exhibitlateral widths releasesaccumulated stress aseismically,numerical cal-
from 300 to 500 km. Larger widths (600-1500km) are culations show that stress can still be transmitted from
observed for plateaupairs(i.e., SW Japan,Peru). The the subducting plate to the crustal regionsof the over-
scaleof these global fiat slab segmentsmatcheswell riding plate [Spencer,1994]. Increasedinterplatecou-
with the sizespredictedby the buoyancycalculations. pling abovefiat slab regionshas also been suggestedto
Once oceaniclithospherebecomesmucholder (e.g., 80 be responsiblefor thick-skinned deformation in the up-
m.y. old and 100 km thick [Cloos,1993])and denser, per plate, resulting in large-scaleblock-type uplifts as
fiat subduction becomesnearly impossible. in the PampeanasRange [Jordanet al., 1983; Sinalley
et al., 1993] and in the Eastern Cordilleraof Colom-
3.3. Impact of Flat Subduction: Interplate bia (Figuresi and 9). Many workershavesuggesteda
Coupling, Deformation and Thermal Structure similar origin for the eastern Rockies Laramide uplifts
While it has been pointed out by previousworkers [Jordanet al., 1983; Bird, 1984, 1988; Dumitru et al.,
that upperplate seismicityis moreabundantabovefiat 1991].
slabregionsthan overneighboring steepslabsegments Interplate couplingalso appearsto be an important
[Jordan½tal., 1983;Sinalleyet al., 1993],no study factor governingstrain partitioningin casesof oblique
has attempted to quantify this variation. We presenta convergence as suggested by analogmodelingand nu-
statisticalanalysisof the upper plate seismicityalong merical modeling [P'i•ct and Cobbold,1992; Pubellier
the entire lengthof the subductingNazca Plate, from and Cobbold, 1996; Chemenda et al., 2000], yet the
8øNto 44øSin Figure8. This coversthe twobestdocu- mechanism for modifying this parameter remains un-
clear. We have shown that fiat subduction increases
mentedfiat slabregionsin PeruandcentralChileaswell
as two inferred shorter fiat slab segmentsin Ecuador interplate coupling and suggestthat fiat subduction
(Carnegie Ridge)andColombia (ChocoBlock)(Figure will enhance strain partitioning at a position far in-
1). The seismic energyreleased by eachearthquake is board of the trench forearcregion (Figure 10). Two
calculated and summedup in each 10 north-southbox prominentexamplesprovide usefulinsights,the NW
for a distance of 250-800 km from the trench. This Andean margin and SW Japan. In both casescon-
spansfrom west of the volcanicarc (whenpresent), vergenceis 20ø-40¸ with respectto trenchnormal,and
across the central cordillera and in all cases, except major dextral strike-slipmotion occurs.GPS data sug-
the 800 km wide central Andes and Bolivian sub-andes gestdisplacement of the northAndesblockby 1-2cm/a
to thewest), alongthe Dolores-Guyaquil
(wheretheBrazilianShieldisunderthrust megashear
(Figure1) [Kel-
all the way to the undeformedforeland.The histogram loggand Vega,1995]. In Ecuadorthis dextral slip (see
analysisof the seismicitydemonstratesthat the seismic focal mechanismsin Figure 9) occurs300-400km from
energyreleasedin the upperplate abovefiat slabseg- the trench, at the level of the adakitic volcanic arc
mentsis on average3-5 times greaterthan in adjacent and appearsto be causedby increasedcouplingabove
steep(>30 ø dipping)slab segments.In someextreme the CarnegieRidgefiat slabsegment[Cutschefet al.,
cases,there is more than an order of magnitudediffer- 1999b]. Similarly,beneathSW Japanfiat subduction
9oow 75ow
10øN
Colom. F.S.
5.4 / 8.8
6.913.7
1.218.8
5.4! 4.1
1.510.
9
Carn. F.S.
oß 5.4 ! 10.3
4.3113.3
Peru
flat slab
8.5 ! 2.2
10øS
Seismic Energy Release
1.0 / 9.0
all flat slab (1964-1995) 7.5
0.7 / 0.0
ß 2.311.4
per hundred years
•..
ß
':•:'"•G
1.813.2...i
::ili::;!';•;i;
for all steep slab segments: 4.8
for all flat slab segments: 13.0
•.C1.911.6
I'"• Central Chile
flat slab
7.9 ! 5.4
30øS
. .... .' ....".•i
•..:
:..:':.
:.:i'
•:":.'
"..i•!::i!iii•iii:•::•i•i::i!i•i!i•:!::ii:•
::i
5.3 / 8.4
1.415.1
SISRA
Catalog
(1900-1963)
0.010.8
: 0.010.8
EHBCatalog
(1964-1995)
..
%
•"' 0.310.3
all energy in units of
......
'""::
........
--• 0•210.1
J/ 100kmmargin
length .. -:....,.,
'........
.7 SChile
40øS•o . ß o.•lO.1 steep slab
o.s/o.o 0.7 / 1.4
0.310.1
0.010.0
0.010.0
.e ...
Figure 9. Relief and focal mechanismsin South America from the Harvard CMT catalog
for shallow(<70 kin) earthquakes(1976-1999). Shadedhill relief from digital topographyand
bathymetrydatabase[Smithand Sandwell,1997].
A)
oceanic lithosphere volcani
"•arc deformation
incordillera
O- accr.wedge • ••
.... ''::'-• • margin ....
.... •on --_ -•4•:-.•.o i warm,aselsmlccrust
ß
--600oc
.............
•'.'--"'L
.L-•.•.
_....;
.... Moho_ • "-..... •o
..........
'- 900ø0
........... '...... "•-•.':.,••_. ...... :-----
-----
.... :--:--=--:-
...........--...........600C..................
.... ..... '........ ...... .............................................
9oooc ..................
'. .... '":
•:'--5:;:'--,...
"'-....'_'-...."
•,';-..•_:" ? continental
lithosphere
[km]
_-
100 -
150 - -* 1.4000C
.., :-;:;:;;",::?•
............ ............
,.......... "?'"".?½:
"'%•'i•i::?:!•i:;:i:i•!:?i
"::':•?';':;!i'i:i-i•i:i'?!;::•!•:;;•;:i•?!:;::
""'::%":;;;:.i;:;:ii•.:::':"?:"?;""!:!?--i•i•!•;:;;;?:'i?,,:..
;•,,½•,'-:>.,',,",.::"::::;;.;i•i-i;;;;;;;;•;;:;-i;;:.;::?!?,;•;•:.::i;•,:7:,
'"'"":•';•;•'•::•":•,'"'"'•'
.......................
:'-•"'
•'•':":':
"::"'•:'
..::":::";:.
•......... " .......
..•......'..}:::.;:-.
!:::i::.::;i::::.:..:.-ii.i-!;::-iii;i'::".::;'
..::::.•:•::i;5:;:(:.:::::::.i;;•;;•:i:::::;.::::i:i..:::.::•;:::;;::•.;.!::.::.i!i.•::•.•*.-•
:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
', ,, c o• ============================================================================
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
•.:.•'-:..:.:,.-;
•.i. ;::..::
.;::.•
:::•.::;.•.;•;•:::::::•.•:•:•:::::::•.•.•:}•.;•.•;.;.;;;::;:f•;;..•.•;.;•:.:.•;•;•.;..;.;:::•;;;;;•;•.:•.;:•::•.;•.:.•:::•;.;:i:;;•;;.;•.;•:;:;•;•.•;;•;•;•;;::•:;•;;•:•.•;•;;::;•.•;•.;•;•;•;;;i;•;;;•;•;;:.:•:•:;:•;i;;;;;.;•;•;•.•;;;;;;
'% ', ,,:,x,?•::•::•i:;iii•
ii::!:;•D.%ii!::.;;•ii!•iiii::!:i1•!!•.•:•.i::::;•;•.•:..•.•!•!•::.•;•E•.::::•!•:.iii•:i
•;.i.i.i::::ii::.!:.i•!;!.i;;:.•.•Ei..!•!•..•::;.!•.!!!.::.!;:i.ii:::5.!.i:i.:.ii:
::::::::::::::::::::::::
i:::
:: ...
I I I I I I I I I ........
i.............................
I.............. I i.............
0 1O0 200 300 [km] 400 500 600 700
...60006....
•..•-'•'%•;-•!•••%,•_._._. ....,
•"cold,
•ei•mog:•nic
crust•,•' ø•' (•'/9
•3 oøc ...... '".... 6øøød'
,.,....::-...0.
•;-.....-
..-
............
;;;:::.'"'-
'••••.' cont,
nental
I,thosphere ....... 9oo0
6.......
50--
[km]
- ':i!:?
':i';i';;
;';';
......
•':::;!:';:;':';•?i;i:,;:;;:!i!:!:;i:;i::i:i.
•i•?i•:i!;•;i•;•;ii;.•!!:?•i!•}ii:.?•....•_.:•....•i;;:;.•.•;?.;:``````•```````
,..•.;•(r•;;;;d
•..¾...
'.'..'. interplate
coupling,-....... 130000
..
..'..... ...;:.;:'.'.;;:;:i::.;:'
i;i';
•;.:;ii(::•..•.:!:•!•:•!:.:•::;:ii!•;:.!.;!•if.!;(:;::•!ii•.:•:...:•.i;5•i•:i•5•i:i!•!i.;!:•:(:i!•!.:::7:.•:::.•::.!.;.•:%
' - ,, '% ".•i.":." ½.• ' •?.:.
•' ' .'•...-'• •; ::•;;::(ii•'::::•5•½-•::::•;:.---.-'-'-.;
-½'
-::"•-::::-;•;;•-•:•:•-•':
100 --
150 --
ß: •'::•'::"'•':•-:
.........!!-i:5;'•?:•-t:::...;,,•-;::.;:i:½2i:;?::i;ii?'.::'
;:•;;.:.•::;:;'-""•
:::-,';ili:-i:;;;:;;?:.:?•;;i;ii!:?;:::i!';;';iii:;:;:::';
............
:::i;½ ;i•::•:;!ii:•;;:i•;i•;:i;?;;•7i•i.i::•?:ii•:•;:•;;•i:;i!:;•i;!:i!•.;:!•....:i.!:i:•.;
:;•'::!::i;;:'•.•.;
..........
;:i?f:-'.!..i;!:;,...-i:!-:';?:'.i:i•..?
..... ?:?,;'.:.ii;:•'•.:i:':!•:
.......
%:'ii!i::•i!:::::.::ii•:?
, •:;:/i%•.',
', ', ",x%;;' .
i i I I I I I I I I I i 1
0 1O0 200 300 [km] 400 500 600 700
Last, we wish to emphasizethe impact of fiat sub- of this asthenosphericwedge which is responsiblefor the
ductionon the temperatureandrheologyof the margin cessationof calco-alkalinearc magmatism[McGeary et
sincethis directly controlsthe deformationprocesses al., 1985; Kay and A bbruzzi,1996]. Thus, prolonged
in the upper plate and along the interface. Flat sub- fiat subduction cools both plates and greatly increases
duction completely alters the thermal structure of the the strengthof the upper plate [Vlaar, 1983; Spencer,
margin, bringingcold oceaniclithosphereto a position 1994]. Finally, the impact of the coldthermal structure
beneaththe upper lithosphere,wheretypically hot as- of fiat subduction on the downdip limit of the seismo-
thenosphereis present,at depths of 60-100 km and 200- genic zone, commonly consideredto be related to the
500 km from the trench(Figure10). The thermaleffect 350ø or 450øCisotherms[Oleskevich et al., 1999],re-
is best illustrated by heat flow observationsfrom Peru mains to be investigated and is potentially of tremen-
andBolivia.Highheatflowof 50-120mW/m • ispresent dous importance for the assessmentof seismic risk.
abovethe steepslabin Boliviaowingto the underlying
asthenosphere,
whereasabovethe Peru fiat slab, where 4. Conclusions
the asthenosphereis absent,low "forearc"valuesof 30-
70 mW/m • areobserved
up to 500km fromthe trench On the basis of the evidence presented here we con-
[HenryandPollack,1988].It is alsothe disappearance cludethe following. (1) Flat subductionis causedpri-
GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION 829
Appendix- Identification of Flat Slab In Costa Rica, 14-20 m.y. old oceanic flanks the sub-
Segments as Supplement to Table 1 ductingCocosRidge [Pvottiet al., 1994]. On the basis
of local network hypocenterdata a general decreasein
A1. Central Chile
slab dip from >600 in Nicaragua and NE Costa Rica to
Togetherwith Peru this wasoneof the first segments 300 in central Costa Rica is reported. In the segment
to be identified as ::low-anglesubduction"[Barazangi facing the CocosRidge, hypocentersdo not extend be-
and Isacks, 1976]. Subsequentwork refined the slab low60 km (justasin SW Japap.or Cascadia),
yet these
geometry,identifying three inflectionpoints, and thus authorsproposethe slabgeometryfiattens,"continuing
redefiningit as "fiat," whichwasattributedto the sub- along an almost horizontal path like tha• found under
duction of the Juan Fernandez Ridge and expressedat PeruandsouthernMexico"[Pvottiet'al., 1994,p. 282].
the volcanic arc by a change in geochemistryfollowed Here as well, there is a gap in Quaternary stratovol-
by a cessationof activity (volcanicgap) [Pilger,1981; canoes despite active subduction and the presenceof
recent ariakites.
Cahill and Isacks,1992;Kay and Abbvuzzi,1996].
The Perusegmentwasfirst identifiednearly25 years This margin was the third to be officially labeled as
ago as low-anglesubduction[Bavazangiand Isacks, fiat subduction[Suarezet al., 1990].The subhorizontal
1976]basedon teleseismicearthquakehypocenters.Lo- slab is unusually shallow, at a depth of 50 km. Inter-
cal network data demonstrated the subduction angle is estingly,in Mexicothe youngest (5-10m.y. old) part of
not constantand shallow,but rather moderately steep the slab (at 102ø-105øW longitude)dipsfairly steeply
300 down to 100 km and then practically horizontal for (>30ø) and is well constrainedby hypoccnters and by
severalhundredkilometers[Hasegawa andSacks,1981] the narrow trench arc gap. As lithosphericagesin-
and first establishedthe conceptof a "fiat slab" geom- creaseto the east, paradoxically,the slab dip decreases
etry. and becomes
truly fiat (at 99ø-96øWlongitude)[Singh
830 GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION
References
Abbott, D., R. Drury, and W.H.F. Smith, ma Basin and Galapagos "Plume": New 96, edited by J. Bialas, and N. Kukowski,
Flat to steep transition in subduction Investigationsof Intraplate Magmatism), 490 pp., GEOMAI• Res. Cent., Univ.
style, Geology, 22, 937-940, 1994. in GEOMAR Report 94, edited by J. Kiel, Kiel, Germany, 2000.
Barazangi, M., and B. Isacks, Spatial dis- Bialas, E. Flueh, and G. Bohrmann, 437 Bijwaard, H., W. Spakman, and E.l•. En-
tribution of earthquakes and subduction pp., GEOMAl• Res. Cent., Univ. Kiel,
gdahl, Closing the gap between regional
of the Nazca Plate beneath South Amer- Kiel, Germany, 1999.
and global travel time tomography, J.
ica, Geology, J, 686-692, 1976. Bialas, J., N. Kukowski, and GEOPECO
Geophys. Res., 105, 30,055-30,078, 1998.
Bevis, M., The curvature of Wadati-Benioff Scientific Team, FS SONNE Cruise Re-
port SO146/1 and 2, GEOPECO (Geo- Bird, P., Laramide crustal thickening event
zones and the torsional rigidity of sub-
physical Experiments at the Peruvian in the Rocky Mountain foreland and
ducting plates, Nature, 323, 52-53, 1986.
Bialas, J., E.R. Flueh, and PAGANINI Continental Margin: Investigations of Great Plains, Tectonics, 3, 741-758, 1984.
Scientific Team, FS SONNE Cruise Re- Tectonics, Mechanics, Gashydrates, and Bird, P., Formation of the Rocky Moun-
port SO144/1 and 2, PAGANINI (Pana- Fluid Transport), in GEOMAR Report tains, western United States: A con-
GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION 831
tinuum computer model, Science, 239, Feigonson,R.C. Maury, and R.H. Stew- tonicsegmentation
of the N. Andes,Ge-
1501-1507, 1988. art, The geochemistryof young volcan- ology, 13, 203-207, 1985.
Blackwell, D.D., J.L. Steele, S. Kelley, and ism throughout western Panama and SE Hardy, N.C., Tectonic evolution of the east-
M.A. Korosec, Heat flow in the state of Costa Rica: An overview, J. Geol. Soc. ernmost Panama Basin: Some new data
Washington and thermal conditions in London, 149, 569-579, 1993. and inferences,
J. S. Am. Earth Sci., 4,
the Cascadia Range, J. Geophys. Res., DeMets, C., R.G. Gordon, D.F. Angus, 261-269, 1991.
95, 19495-19516, 1990. and C. Stein, Current plate motions, Hasegawa,A., and I.S. Sacks,Subduction
Boyd, T.M., J.A. Snoke, I.S. Sacks, and Geophys. J. Int., 101, 425-478, 1990. of the Nazca Plate beneath Peru as de-
A. Rodriguez, High-resolution determi- Dumitru, T.A., P.B. Gans, D.A. Foster, termined from seismicobservations,J.
nation of the Benloft Zone geometry be- and E.L. Miller, Refrigerationof the west- Geophys.Res., 86, 4971-4980,1981.
neath Southern Peru, Bull. Seisin. Soc. ern Cordilleranlithosphereduring Lara- Henry, S.G., and H.N. Pollack, Terrestrial
Am., 74, 559-568, 1984. mide shallow-anglesubduction,Geology, heat flow above the Andean subduction
Brocher, T., G.S. Fuis, M.A. Fisher, G. 19, 1145-1148, 1991. zone in Bolivia and Peru, J. Geophys.
Plafker, and M.J. Moses, Mapping the Eberhart-Phillips, D., and M. Reyners,Con- Res., 93, 15,153-15,162, 1988.
megathrust beneath the northern Gulf tinental subduction and three dimen- Hey, R.N., Tectonic evolution of the Cocos-
of Alaska using wide-angle seismicdata, sional crustal structure: The northern Nazca spreadingcenter, Geol. Soc. Am.
J. Geophys. Res., 99, 11,663-11,685, South Island, New Zealand, J. Geophys. Bull., 88, 1414-1420, 1977.
1994.
Res., 102, 11,843-11,861, 1997. Hirahara, K., Three-dimensional seismic
Cahill, T., and B. Isacks, Seismicity and Engdahl, E.R., R.D. van der Hilst, and structure beneathsouthwestJapan: The
shape of the subducted Nazca Plate, J. J. Berrocal, Imaging of subductedlitho- subductingPhilippineSeaplate, Tectono-
Geophys. Res., 97, 17,503-17,529, 1992. spherebeneathSouthAmerica, Geophys. physics, 79, 1-44, 1981.
Cande, S.C., and W.F. Haxby, Eocene prop- Res. Lett., 22, 2317-2320, 1995. Isacks, B., Uplift of the central Andean
agating rifts in the southwest Pacific and Engdahl, E.R., R.D. van der Hilst, and R. plateau and bendingof the Bolivian oro-
their conjugate features on the Nazca Buland, Global teleseismic earthquake cline, J. Geophys.Res., 93, 3211-3231,
Plate, J. Geophys. Res., 96, 12,061- relocation with improved travel times 1988.
12,084, 1991. andprocedures for depthrelocation,Bull. James,D.E., and J.A. Snoke,Structure
Caress, D.W., M. McNutt, R.S. Detrick, Seismol. Soc. Am., 88, 722-743, 1998. and tectonicsin the regionof flat sub-
and J.C. Mutter, Seismicimaging of hot- Feighner, M.A., and M.A. Richards, Litho- duction beneath central Peru: Crust and
spot related crustal underplating beneath sphericstructure and compensationmech- uppermost mantle, J. Geophys. Res.,
the Marquesas Islands, Nature, 373, 600- anismsof the GalapagosArchipelago,J. 99, 6899-6912, 1994.
603, 1995. Geophys. Res., 99, 6711-6729, 1994. Jischke,M., On the dynamics of descend-
Charvis, P., M. Recq, S. Operto, and D. Flueh, E.R., et al., New seismicimages ing lithospheric
platesand slipzones,J.
Brefort, Deep structure of the northern of the Cascadia subduction zone from Geophys.Res., 80, 4809-4813, 1975.
Kerguelen Plateau and hotspot-related cruiseSO108-ORWELL,Tectonophysics,
Jordan,T., B. Isacks,R.W. Allmendinger,
activity, Geophys. J. Int., 122, 899-924, 293, 69-84, 1998. J.A. Brewer,V.A. Ramos,andC.J. Ando,
1995.
Gardner, T.W., D. Verdonck, N.M. Pin- Andeantectonicsrelatedto geometryof
Chemenda, A., S. Lallemand, and A. Bokun, ter, R. Slingerland,K.P. Furlong, T.F. subducted Nazca plate, Geol. Soc. Am.
Strain partitioning and interplate fric- Bullard, and S.G. Wells, Quaternary up- Bull., 94, 341-361, 1983.
tion in oblique subduction zones: Con- lift astride the aseismicCocos Ridge, Kay, S.M., and J.M. Abbruzzi,Magmatic
straints provided by physical modeling, Pacific coast, Costa Rica, Geol. Soc. evidencefor Neogenelithosphericevo-
J. Geophys. Res., 105, 5567-5582, 2000. Am. Bull., 104, 219-232, 1992. lution of the central Andean "fiat-slab"
Cloos, M., Lithospheric buoyancy and col- Gephart, J.W., Topography and subduc- between30øSand 32øS,Tectonophysics,
lisional orogenesis:Subduction of oceanic tion geometry in the central Andes: clues 259, 15-28, 1996.
plateaus, continental margins, island arcs, to the mechanics of a noncollisional oro- Kellogg,J.N., and V. Vega, Tectonicde-
spreading ridges, and seamounts, Geol. gen, J. Geophys. Res., 99, 12,279-12,288, velopmentof Panama, Costa Rica, and
Soc. Am. Bull., 105, 715-737, 1993. 1994. the Colombian Andes: Constraints from
Cross, T.A., and R.H. Pilger, Controls of Gutscher, M.-A., An Andean model of in- GlobalPositioning System(GPS)geode-
subduction geometry, location of mag- terplate coupling and strain partition- tic studiesand gravity, in Geologicand
matic arcs, and tectonics of arc and back- ing applied to the fiat subduction of Tectonic Development of the Caribbean
arc regions, Geol. Soc. Am. Bull., 93, SW Japan (Nankai Trough), Tectono- Plate Boundary in Southern Central Amer
545-562, 1982. physics, in press, 2000. ica, edited by P. Mann, Spec. Pap.
Crosson,R.S., and T.J. Owens,Slab geom- Gutscher,M.-A., and S. Lallemand,Birth Geol. Soc. Am., 295, 75-90, 1995.
etry of the Cascadia subduction zone be- ofa majorstrike-slip
faultin SW Japan, Kodaira,S., N. Takahashi,
J.-O. Park,K.
neath Washington from earthquake hypo- Terra Nova, 11, 203-209, 1999. Mochizuki,M. Shinohara,andS. Kimura,
centers and teleseismic converted waves, Gutscher, M.-A., J.-L. Olivet, D. Asla- WesternNankaiTroughseismogenic zone:
Geophys. Res. Lett., 14, 824-827, 1987. nian, R. Maury, and J.-P. Elssen,The Resultsfrom a wide-angleoceanbottom
Davy, B., and R. Wood, Gravity and mag- 'lost Inca Plateau': Cause of fiat sub- seismicsurvey, J. Geophys.Res., 105,
netic modeling of the Hikurangi Plateau, duction beneath Peru?, Earth Planet. 5887-5905, 2000.
Mar Geol., 118, 139-151, 1994. Sci. Lett., 171, 335-341, 1999a. Kolarsky,R.A., P. Mann,andW. Montero,
Derant, M.J., and M.S. Drummond, Deriva- Gutscher,M.-A., J. Malavieille, S. Lalle- Island arc response to shallow subduc-
tion of some modern arc magmas by mand, and J.-Y. Collot, Tectonicseg- tion of the CocosRidge, Costa Rica, in
melting of young subducted lithosphere, mentationof the North AndeanMargin: Geologicand TectonicDevelopmentof
Nature, 347, 662-665, 1990. Impact of the CarnegieRidge collision, the CaribbeanPlate Boundary in South-
Derant, M.J., and M.S. Drummond, Mount Earth Planet. Sci. Lett., 168, 255-270, ern CentralAmerica, editedby P. Mann,
St. Helens: Potential example of the 1999b.
Spec. Pap. Geol. Soc. Am., 295, 235-
partial melting of the subducted litho- Gutscher, M.-A., E. Bourdon, J.-P. Eis- 262, 1995.
sphere in a volcanic arc, Geology, 21, sen,and R. Maury,Can slabmeltingbe Le Pichon,X., S. Mazzotti, P. Henry,and
547-550, 1993. causedby fiat subduction?,
Geology,28, M. Hashimoto,Deformationof the Japa-
Defant, M.J., T.E. Jackson, M.S. Drum- 535-538,2000. neseIslandsand seismiccoupling:an
mond, J.Z. de Boer, H. Bellon, M.D. Hall, M.L., and C.A. Wood, Volcanotee- interpretationbaseelon GSI permanent
832 GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION
GPS observations, Geophys. J. Int., subduction, Tectonophysics, 102, 377- try and ship depth soundings, Science,
184, 501-514, 1998. 398, 1984. 277, 1956-1962, 1997.
Lonsdale, P.. and K.D. Klitgord, Struc- Pilger, R.H., Plate reconstructions, aseis- Spencer, i.E., A numerical assessment of
ture and tectonic history of the eastern mic ridges, and low-angle subduction slab strength during high- and low-angle
Panama Basin, Geol. Soc. Am. Bull., beneath the Andes, Geol. Soc. Am. subduction and implications for Lara-
89, 981-999, 1978. Bull., 92, 448-456, 1981. mide orogenesis, J. Geophys. Res., 99,
McCann, W.R., and R.E. Habermann, Mot- Pilger, R.H., Cenozoic plate kinematics, 9227-9236, 1994.
phological and geological effects of the subductionand magmatism: South Amer- Suaxez,G., P. Molnar, and B.C. Burchfiel,
subduction of bathymetric highs, in Sub- ican Andes, J. Geol. Soc. London, 141, Seismicity, fault plane solutions, depth
duction Zones, edited by L.J. Ruff and 793-802, i 984. of faulting, and active tectonics of the
H. Kanamori, pp. 41-69, Birkh•user Pinet, N., and P.R. Cobbold, Experimen- Andes of Peru, Ecuador and S. Colom-
Boston, Cambridge, Mass., 1989. tal insights into the partitioning of mo- bia, J. Geophys. Res., 88, 10403-10428,
McGeary, S., A. Nut, and Z. Ben-Avraham, tion within zones of oblique subduction, 1983.
Spatial gaps in arc volcanism: The ef- Tectonophysics, 206, 371-388, 1992. Suarez, G., T. Montfret, G. Wittlinger,
fect of collision or subduction of oceanic Protti, M., F. Guendel, and K. MeNally, and C. David, Geometry of subduction
plateaus, Tectonophysics, 119, 195-221, The geometry of the Wadati-Benioff zone and depth of the seismogeniczone in the
1985. under southern Central America and its Guerrero gap, Mexico, Nature, 345, 336-
Monzier, M., C. Robin, M.L. Hall, J. Cot- tectonic significance: Results from a high- 338, 1990.
ten, P. Mothes, J.-P. Eissen, and P. Sama- resolution local seismographic network, Talandier, J, and E. Okal, Crustal struc-
niego, Les adakites d'Equateur: modele Phys. Earth Planet. Inter., 84, 271- ture in the Society and Tuamotu Is-
pr•liminaire, C. R. Acad. Sci., Set. I, 287, 1994. lands, French Polynesia, Geophys. J. R.
$2J, 545-552, 1997. Protti, M., F. Guendel, and K. MeNally, Astr. Soc., 88, 499-528, 1987.
Morris, P.A., Slab melting as an explana- Correlation between the age of the sub- Uyeda, S., and H. Kanamori, Back-arc
tion of Quaternary volcanism and aseis- ducting Cocosplate and the geometry of opening and the mode of subduction, J.
micity in southwest Japan, Geology, 25, the Wadati-Benioff zone under Nicaragua Geophys. Res., 84, 1049-1061, 1979.
395-398, 1995. and Costa Rica, in Geologic and Tec- van der Hilst, R., apd P. Mann, Tectonic
Nakanishi, I., Precursorsto ScSphasesand tonic Development of the Caribbean Plate implications of tomographic images of
dipping interface in the upper mantle Boundary in Southern Central Amer- subducted lithosphere beneath northwest-
beneath southwestern Japan, Tectono- ica, edited by P. Mann, Spec. Pap. ern South America, Geology, 22, 451-
physics, 69, 1-35, 1980. Geol. Soc. Am., 295, 309-326, 1995. 454, 1994.
Nut, A., and Z. Ben-Avraham, Volcanic Pubellier, M., and P. Cobbold, Analogue van der Velden, A.J., and F.A. Cook, Pro-
gaps and the consumption of aseismic models for the transpressional docking terozoic and Cenozoic subduction com-
ridges in South America, in Nazca Plate: plexes: A comparison of geometric fea-
of volcanic arcs in the Western Pacific,
Crustal Formation and Andean Conver- Tectonophysics, 253, 33-52, 1996. tures, Tectonics, 18, 575-581, 1999.
gence, edited by L. Kulm et al., pp. 729- Pubellier,M., G. Prouteau,R.C. Maury, Vlaar, N.J., Thermal anomalies and mag-
740, Geol. Soc. Am., Boulder, Colo., andH. Bellon,Le changement der•gime matism due to lithospheric doubling and
1981.
plio-quaternairedansunezoned'dchappe- shifting, Earth Planet. Sci. Lett., 65,
Okal, E.A., Historical seismicity and seis- ment tectonique; controle par des in- 322-330, 1983.
moteetonic context of the Great 1979 Vlaar, N.J., Precambrian geodynamical con-
trusifs syntectoniques
(Irian Jaya, In-
Yapen and 1996 Biak, Irian Jaya earth- don•sie),Proc. RduniondesSciencesde straints, in The deep Proterozoic crust
quakes, Pure Appl. Geophys., 154, 633- la Terre (RST Meeting)Brest,France, in North Atlantic Provinces, edited by
675, 1999. 179-180, 1998. A.C. Tobi and J.L.R. Touret, pp. 3-20,
Olbertz, D., M.J.R. Wortel, and U. Hansen, Reidel Publ. Co., 1985.
Sacks,I.S., The subductionof younglitho-
Trench migration and subduction zone sphere, J. Geophys. Res., 88, 3355- Vlaar, N.J., and M.J.R. Wortel, Litho-
geometry, Geophys. Res. Lett., 24, 221- 3366, 1983. spheric aging, instability and subduc-
224, 1997. tion, Tectonophysics, 32, 331-351, 1976.
Singh, S.K., and F. Mortera, Source time
Oleskevich, D.A., R.D. Hyndman, and K. yon Huene, P•., I. Pecher, and M. Gutschef,
functions c{ large Mexican subduction
Wang, The updip and downdip limits earthquakes, morphology of the Benioff Development of the accrctionary prism
to great subduction earthquakes: Ther- Zone, age of the plate and their tec- along Peru and material flux after sub-
mal and structural models of Cascadia, duction of Nazca Ridge, Tectonics, 15,
tonic implications, J. Geophys. Res.,
south Alaska, SW Japan, and Chile, J. 96, 21487-21502, 1991. 19-33, 1996.
Geophys. Res., 104, 14,965-14,991, 1999.
Parsons, T., A.M. Trehu, J.H. Luetgert,
Smalley,R., and B. Isacks,A high reso- yon Huene, R., J. Corvalan, E.R. Flueh,
lution local network study of the Nazca K. Hinz, J. Korstgard, C.R. Ranero,
K. Miller, F. Kilbride, R.E. Wells, M.A.
Plate Wadati-Benioff Zone under West- W. Weinrebe and the Condor Scientists,
Fisher, E. Flueh, U.S. ten Brink, and
N.I. Christensen, A new view into the ern Argentina, J. Geophys. Res., 92, Tectonic control of the subducting Juan
Cascadia subduction zone and volcanic
13903-13912, 1987. Fernandez Ridge on the Andean margin
Sinalley,R., J. Pujol, M. Regnier, J.-M. near Valparaiso, Chile, Tectonics, 16,
arc: Implications for earthquake haz-
Chiu, J.-L. Chatelain, B.L. Isacks, M. 474-488, 1997.
ards along the Washington margin, Ge-
ology, 26, 199-202, 1998. Araujo, and N. Puebla, Basement seis- Wessel, P., and W.H.F. Smith, Free soft-
Peacock, S.M., T. Rushmet, and A. Thomp- micity beneath the Andean Precordillera ware helps map and display data, Eos
son, Partial melting of subducting oceanic thin skinned thrust belt and implica- Trans. A GU, 72 (41), 441, 1991.
crust, Earth Planet. Sci. Lett., 121, tions for crustal and lithosphericbehav- Wortel, M.J.R., Spatial and temporal vari-
227-244, 1994. ior, Tectonics, 12, 63-76, 1993. ations in the Andean subduction zone,
Pennington, W.D., Subduction of the east- Smith, W.H.F., and D.T. Sandwell, Bathy- J. Geol. Soc. London, 141, 783-791,
ern Panama Basin and seismotectonics metric prediction from dense altimetry 1984.
of northwestern South America, J. Geo- and sparseshipboardbathymetry,J. Geo- Wortel, M.J.R., and N.J. Vlaar, Age de-
phys. Res., 86, 10,753-10,770, 1981. phys. Res., 99, 21803-21824, 1994. pendent subductionof oceaniclithosphere
Pennington, W.D., The effect of oceanic Smith, W.H.F., and D.T. Sandwell,Global beneath western South America, Phys.
crustal structure on phase changes and seafloortopography from satellite altime- Earth Planet. Inter., 17, 201-208, 1978.
GUTSCHER ET AL.: GEODYNAMICS OF FLAT SUBDUCTION 833
H. Bijwaard and W. Spaklnan, Vening University of Colorado, Boulder, CO, 29280, Plouzand, France. (gutscher@mail-
Meinesz Research School of Geodynamics, sdt.univ-brest.fr)
80309.(engdahl@armstrong.colorado.edu)
Faculty of Earth Sciences,Utrecht Uni- M.-A. Gutscher, Universit• de Bretagne
versity, NL-3508, Utrecht, Netherlands. Occidentale / Institut Universitaire Eu- (receivedJuly 6, 1999;
copden de la Mer, UMR 6538 Domaines revised May 2, 2000;
(wims@geo.uu.nl)
E.R. Engdahl,Department of Physics, Ocdaniques,Place Nicolas Copernic, F- accepted May, 2, 2000.)