Você está na página 1de 17

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 93, NO.

B3, PAGES 2141-2157, MARCH 10, 1988

Age andSpreadingHistoryof The Caynan TroughasDeterminedFromDepth,


Heat Flow, andMagneticAnomalies

ERIC ROSENCRANTZ, MALCOLM I. ROSS, AND JOHN G. SCLATER

Institutefor Geophysics,
Universityof Texasat Austin

The openingof the CaymanTroughrecordsaspectsof relativemotionbetweenthe North AmericanandCaribbean


platesessentialto accuratereconstruction of Caribbeanevolution. Assessment of openingratesbasedupon
basement subsidence analysisaswell asmagneticanomalyinterpretation showsthatthetroughmayhaveopenedat
onethirdto onehalfpreviously proposed rates.Depthsto basement showthattheernsthassubsided awayfromthe
presentaxis in two stages:a youngerstagecharacterizedby substantialand asymmetric(greaterto the east)
subsidence, preceded by an olderstageof slightincreases
of depthwith distance.Depthto agerelationscalculated
from subsidence curveswhichaccommodate lateralaswell asverticalcoolingof the aecretingslabindicatethatthe
troughhasopenedat an overallrateof lessthan 15 mm/yr since25-30 Ma and at 20-30 mm/yr prior to 30 Ma.
Marinemagneticanomalies showlinearions perpendicular to thetrough,includinga distinctcentralanomaly,but
thepatternis neitherwell definednorsymmetrical aboutthepresentaxis. However,partialanomalysequences can
be identifiedonbothsidesof theaxis,andthedistribution of thesesequencesindicatesthatspreading mayhave
involvedfrequentshortridgejumps,mosfiyeastward.Theoverallrateof openingasindicated by thisinterpretation
of themagnetics is about15 and30 mm/yrsinceandpriorto 26 Ma. Bothsubsidence andmagnetics indicatethat
thetroughopenedby at least45-50 Ma.

INTRODUCI•ON connectingthe two transformfaults. The history of relative


motionrecordedin the crustaccretedat thisspreadingcenterboth
Reconstructing the evolutionof the Caribbeanhasan important
outlinesthe ageanddurationof tectoniceventsalongthenorthern
bearingon understanding globalplateprocesses in severalways.
boundaryof the Caribbeanand providesa measureof constraint
The geologyof theregionrecordsthetectonicinteractionof several
over the relative motionsbetweenthe Caribbeanplate and its
majorplatesthroughtime. This interactionhasprimarilybeen
surrounding plates.
thatof theNorth andSouthAmericanplatesbut alsoincludesthat
Marine magneticanomaliesin the CaymanTroughhave been
of severalPacificplateswith the North America-South America
identifiedby MacdonaldandHolcombe[1978]. However,therates
pair. Unravelingthe natureof this interactionbearsnot only on
of spreadingthey computedfrom their identificationsare much
the geologicalhistoriesof thevariousplatesthemselves but also
fasterthanthoseimpliedby crustalsubsidence, and their crustal
on the questionsas to why suchregionsof complextectonics
ages and age of trough openingare much younger than those
developandwhy they developthe way they do. Finally, the impliedby theregionalgeology [Rosencrantzand Sclater, 1986].
sedimentsof the region provide a look at the nature of past
Unfortunately, no samplesclearlydocumentingbasement agehave
connectionsbetweenthePacificandAtlanticoceansystems, which
been collectedfrom within the CaymanTrough,exceptat the
in turnhasanimportantbearingonreconstructing paleoclimates.
presentspreadingcenter,so thereis no independent measureof
One of the major problemsencountered in reconstructingthe
crustal age against which the magneticscan be calibrated.
tectonic
evolution
oftheCaribbean
isthat
itisdifficult
totietheHowever,because
theCayman Spreading
Center
displays
allthe
motion
oftheCaribbean
plate
tothose
ofitssurrounding
plates major
topographic
andlithological
characteristics
ofanoceanic
(North
American,
SouthAmerican,
Nazca,
and Cocos).
Caribbean spreading
center
[Holcombe
etal.,1973;
Macdonald
andHolcombe,
plate
boundaries
havebeenpredominantly
either
strike
slipor 1978;
CAYTROUGH, 1979],itisreasonable
toassume
thatits
convergent
(Figure
1)and
contain
atbest
asubtle
record
ofrelative
thermal
characteristics
shouldalso
beoceanic.
Thesubsidence
platemotion. Consequently,
anystructure
withintheseboundary
versusageandheatflow versusagerelationships documented for
zoneswhich does provide good indicationsof relative motion
crustwithin the deepoceans[e.g.,Sclateret al., 1971;Davis and
becomesanimportant
factor
inregional
reconstructions. Lister, 1974;Parsons
andSclater,
1977;
Lister,
1977]
andthose
TheCayman Troughisoneofthesecritical
structures.
Thedeveloped
byBoerruer
andSclater
[1987]
forsmall
isolated
ocean
trough
marksthe
present
position
ofthe
northern
Caribbeanstrike-
basins
should
apply
tothe
CaymanSpreading
Center.
slipplate
boundary
[Molnar andSykes,
1969],
definedbythe Weexamine the
spreading
history
ofthe
Cayman Trough
byfirst
Odeme andSwanIsland
transform
faults.
Holcombeetal.[1973]establishing
estimates
ofcrustal
ageandratesofopening
through
showed
thattopography
central
tothetrough
wassymmetrical
ananalysis
of crustal
subsidence
andheatflow. Wethen
aboutaN-Strending,
flat-floored
valley
atlongitude
81ø40'W
andreexamine
magnetic
anomaly
sequences
inlight
ofthese
newages
thatthisvalley
markedthelocation
of a spreading
centerand
rates
todetermine
trough
spreading
lfistory.
Weshow
that
the
magneticanomalypatternmay be interpretedsuch that it is
consistent
withcrustalsubsidence.
Ourinterpretation
suggests that
Copyright
1988
bythe
American
Geophysical
Union. thetrough
has
opened
intwostages,
anearly
stage
ofrelatively
fastopeningfollowedby a laterandpresentstageof very slow
Paper
number
7B6039. opening,
andthatits opening historyis markedby several
ridge
0148-0227/88/007B-6039505.00 jumps.

2141
2142 ROSENCRANTZ
ETAL.: AGEANDSPREADING
HISTORYOFTI-IECAYMANTROUGH

Honduras CARIBBEAN
Jo

'"

__ PACiFiC OCEAN

Fig.1. Sketch
mapshowing
themajortectonic
features
of theCaribbean
region,
withthelocation
of theCayman
Trough
outlined.
Abbreviations
areasfollows:
Jo,Jocotan-Chameleon
fault;Mo,Motagua
fault;andPo,Polochic
fault.

GEOLOGICAL
ANDGEOPHYSlCAL
OVERVIEW withrugged
topography
consisting
of N-Strending
ridges
and
troughswith maximum reliefs of over 2000 m. The easternand
TheCayman
Troughis anelongate,
rectangular
E-Wtrendingwestern
endsof thetroughhavewaterdepths
whichexceed
4500
deepbasin
which
liessouth
of theCayman RidgeandCubaand M andareunderlain
bysediments
upto1.5sthick(two-waytravel
north
ofHonduras,
theNicaragua
Rise,andJamaica
(Figure
2). It time).Thepresent
position
ofthenorthern
Caribbeanstrike-slip
is bounded northandsouthby precipitous
escarpments which plateboundary
isdefined bytheOrienteandSwan Island
transform
locally
risemorethan5000rnfromthebasin floor.Thecentralfaults [Molnar
andSykes, 1969].Holcombeetal.[1973]identified
thirdofthetrough(between
longitudes
80øand84øW) isflooreda N-Strending, flat-floored
valleyatlongitude
81ø40'W asthe

88 87 86 85 84 8; 81 ORIENTETRANSFORM
FAI 77 76

19-

18-
-18

17- -17
ß

16-

SWAN
ISLANDS
TRANSFORM
FAU•
84 83 82 81 80 79 78 77 76

Fig.2. Bathymetryandmajortectonicfeatures
of theCayman Trough.Bathymetric contours arein kilometers,withthe
exception
ofthedashedline,whichdefines
the200-m contour.
Contour
lines
withticsenclose
deeps. Theheavy solidlinetraces
thepresent
NorthAmerican-Caribbean
plate
boundary. Earthquake
epicenters
areshown asdots(bothsolidandopen).Earthquakes
includeevents
between January1963andSeptember 1986,aslistedbytheInternational
Seismological Center(ISC).Focal
mechanismsarebestdouble-couple
solutions
fromtheHarvardcentroid
moment tensorsolutions
[see
Dziewonski etal.,1981]and
areshownasequalangle,lower
hemisphere
projections,
withshaded
quadrants
representing
compressivefirstmotions.
ROSEN• ET AL.: AGE AND SPREADING HISTORY OF THE CAYMAN TROUGH 2143

location
of a spreading
centerconnecting
thetwotransform
faults. fromthewestern
endof thenorthernfaultscarpandthesouthern
Topography central
tothetroughappears
symmetrical
aboutthis walloftheOriente
transform
faultyieldLateCretaceous
K-Arages
spreading
center.This interpretation
agreedwith othergeophysical [Perfitandtleezen,1978].
datapreviously
collected
withinthetrough[MolnarandSykes, Thegeology
of Guatemala,
to thewestof theCayman
Trough,
1969; Ewing et al., 1960;Bowin, 1968,Edgar et al., 1971; yieldsno firm evidence
of CaymanTroughevolution,
otherthan
Ericksonet al., 1972] and was subsequently
provedby detailed that strike-slipmotionhasoccurredbetweenthe North American
mappings
of the area[PerfitandHeezen,1978;CAYTROUGH, andCaribbean
platessincetheOligocene.
Thismotionhasbeen
1979;Stroup
andFox,1981]. accommodated
alongtheeast-west trendingPolochic,
Motagua and
Macdonald
andHolcombe [1978]computed spreading
ratesforthe Jocotan-Chameleon
left-lateral,
strike-slip
faultsystems.
Burkart
Cayman Troughbased upontheidentification
of anomalies
alonga [1983]showed thatthePolochic faultdisplacesthemiddleto late
single
magneticprofilecrossing
thespreadingaxis.TheircalculatedMiocenevolcanic
arcof CentralAmerica by approximately 130
rateswere20 mm/yrfor theperiod0-2.4Ma and40 mm/yrfor kmandthattheeastern endof thePolochic faultterminates
against
2.4-8.5Ma. The youngerrate(20 mm/yr)agreeswith North theMotaguafault. DeatonandBurkart[1984]arguedthatthis
American-Caribbean
instantaneous
relativemotionascalculated
on displacement
tookplacebetween10.3 and6.6 Ma. On the other
thebasisof globalrelativeplatemotions[e.g.,Jordan,1975]. hand,geologicalevidencefrom the regionof Jamaicaand
Sykeset al. [1982]suggestedthatspreadingslowedat 2.4 Ma Hispaniola,
to theeastof thetrough,
indicates
thatit opened
in
because
a new plateboundary
developed
alongthePlantain
Garden- Eocenetime. Land [1979] suggested
thata seriesof horstlike
Enriquillofaultsystem transecting
Jamaica andHispaniola [Mann blocksat theeastern marginof thetroughnorthof Jamaica are
etal., 1983,1984]. cappedwith Eoceneageclasticdeposits whichmay be the
The spreading ratesof Macdonald andHolcombedo not fully downfaulted equivalents of similarnearbyJamaican rocks. In
agreewith severalbits of geologicalinformation whichrelate Hispaniola,
Bizonet al. [1985]attributed
stratigraphic
anomalies
directlytorelativeplatemotion.Holcombe etal. [1973]observedwithinHaitianOligocene andMiocenesediment sequencesto
thatalthough the longitudinal
sectionof theCaymanTroughis crustalmovements relatedto strike-slipfaultingand further
indistinguishable
from thoseof slow spreadingmid-oceanrises, suggested
thattheearlyto middleEocenealkalinevolcanicsseenin
depths
tobasement aresignificantly
greater
thanthoseobserved at thispartof Haitimarktheonsetof thisfaulting.ManyCaribbean
othermid-oceanicspreadingcenters.They appliedsubsidencereconstructions [e.g.,PindellandDewey,1982;Sykeset al., 1982;
curvesderivedfor theNorthPacific[Sclateret al., 1971]to a Burkeet al., 1984] proposea lateEocene(36-40Ma) timeof
smoothedversionof basementtopography.
Thesecurves suggested CaymanTroughopening,basedlargelyon reconstructions of
that the trough opened at an estimated rate of 22 mm/yr. Eoceneplatepositions.
Rosencrantzand Sclater [1986] comparedcrustal depths to
subsidence curves which accommodated lateral as well as vertical
DE• TO BASEMENT
cooling of lithosphere [Boerner and Sclater, 1987]. These
subsidencecurvessuggestthat the troughhas openedat a rate of Holcombeet al. [1973] notedthatbathymetricprofilesof the
only about15 mm/yr for the past30 m.y. CaymanTroughorientedparallelto the lengthof the basinare
Sparsegeological information alsoindicates thatthetroughtypical ofthose acrossdeepoceanspreadingcenters,withbasement
opened at slowrates.Perfit[1977]argued thatmarlylimestonesslopingawayfromthespreading axis.Theprofileswereunusual
containing Miocene-Pliocene fossilfragments dredged nearthe onlyin that waterdepthsweresignificantly deeperthanthose
present spreading axisshowthatthetrough hasopened ata rateof observed at typicalspreading centers[Holcombe et al., 1973;
approximately 10 mm/yrfor thepast5 m.y. AlongtheMotagua CAYTROUGH,1979].Rosencrantz andSclater[1986]examined
faultin Guatemala, westof thetrough, measurements of recent depths to basement ascalculatedfrom theseseismic profiles.
offsetsindicatethatthepresent rateof sliponthatfaultisbetweenTheyshowed thatbasement topographyalongthetrough suggests
1.5and6 mm/yr[Schwartz et al., 1979]. Although equivalentthatit opened in twodistinct
stages.Theyounger stageproduced
amounts of undocumented displacement couldoccuroneitherthe crestcharacterized by hightopographic reliefandappreciable
nearbyPolochicor Jocotan faults,raisingtotalrateof relative subsidence awayfromthepresent spreading axis,whereas crest
displacement to perhapsasmuchas 18mm/yr[Schwartz et al., generated duringthe olderstageshowslesserrelief andsmall
1979],ErdlacandAnderson [1982]argued thatdisplacementonthe increasesindepthwithincreasing distancefromtheaxis.Crustal
Polochic
faultsince
Miocenetimehasbeenatmostakilometer
or subsidencewithinthecentral,
youngerpartofthetrough
indicated
two,andthereisnoevidence
forPresent
movementontheJocotanthattherateof opening
during
thelaterstage
wasabout15mm/yr.
fault. Thiswouldsuggest
thatthemeasured
displacements
along An intensivereexamination
of the distributionof basement
theMotaguasystemmayin factreflecta current,veryslow depthsin thetrough,incorporating
depthscalculated
fromall
Caribbean-North
American
rateof strike-slip
motion
in Centralavailable
seismic
profiles,
including
thefiveprofiles
examined
by
America. Rosencrantzand Scalter[1986],andfrom the multibeam
Macdonald andHolcombe[1978]
notedthatanextrapolation
of bathymetric
mappresented
byCAYTROUGH [1979]
confirms
the
theirspreading
rates
along
thelength
oftheCaymanTroughplacedpriorobservations
ofHolcombe
etal.[1973]andRosencrantz
and
theopening
of thetrough
in theOligocene.
Wadge
andBurke Sclater
[1986]. Thedistribution
of these
depths
tobasement,
[1983]arguedthatthechange
fromshallowtodeep-watercarbonate
shown asrangesandmeans of depthsobserved at 5 kmintervals
depositionalongthenorth(Cayman Ridge)wall of thetrough alongthelengthof thetrough,isdisplayedin Figure3.
[PerfitandHeezen,1978;EmeryandMilliman,
1980]indicates
that Although theoverallpatternof depthrelativeto thespreading
oceancrestfirstdeveloped
withinthetroughat thattime. Other axisis oneof increasing
depthawayfromtheaxis,it is clearthat
geologicalevidenceLndicates
only that openingis olderthan theseincreases donotdescribesmooth,
continuouscurvesalong
Neogene and younger than Cretaceous. Micritic and marly the length of the basin. Basementwithin the central basin,
limestonesof Mioceneandyoungerageshavebeenrecoveredbetween 220 km westof and165km eastof the axis,clearly
withinthe trough,andgranodiorites
andtonalitesdredged
both slopesawayfromtheaxismoresteeplyandshowsgreaterrelief
2144 ROSEN• ET AL.- AGE AND SPREADING HISTORY OF TI-IE CAYMAN TROUGH

0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0

i i I i i i i i i i I i i

--

--

--

ß
ß
ß
ß
ß
ß
ß
ß
ß

ee

0 I 0 I 0 I 0 !
0 0 I 0 I 0 I 0 I 0 I I
0 0 I 0 I J
•"g • ,- 0 0 0 0 0 0 0 0 0 0
0 •"• 0 •"• 0 •"• 0 •"• 0 •"•
•/N31•13•n•v31•l
-I0 •13Bl•lnN •13/31•1 NI H/d30
ROSEN• ET AL.: AGE AND SPREADING HISTORY OF THE CAYMAN TROUGH 2145

TABLE 1. CaymanTroughHeat Flow Measurements


Latitude, Longitude, Milliwatts
Deg Deg Per Square HFU Qualitya
meter

Epp et al. [1972]


18.55 79.45 67.3b 1.61 P
Ericksonet al. [1972]
16.72 86.90 56.0b 1.34 P
17.18 85.95 89.0b 2.13 G
17.77 84.92 72.7b 1.74 E
17.72 83.75 94.4b 2.26 E
18.72 82.27 94.0 2.25 E
18.70 81.85 88.2 2.11 E
18.98 81.17 91.1 2.18 G
18.55 80.53 49.3 1.18 G
19.08 80.53 94.8 2.27 G
a As published.P, F, G, andE indicatepoor,fair, good,andexcellent
measurements,respectively.
b Measurementsfrom well-sedimented crust.

than basementbeneaththe outerpartsof the basin. Basement but is characterizedas being of poor quality. The four western
westof theaxisrisesfromwesttoeastfroma meandepthof about measurements
rangein valuefrom94.4to 56.0mW m-2 andshow
5600 m at 220 km to about3500 m at 40 km. Local basement anoveralltrendtowardlowervalueswith increasing distanceaway
relief within thissegmentrangesfrom about600 to over2000 m, fromthe axis. One of thesemeasurements
is alsoratedasbeingof
with relief increasing
toward the axis. Basementunderlyingthe poorquality.
westerntroughbeyond220 km showslittle progressive changein
depthoutto thelimit of oceaniccrustat 500 km [Rosencrantz
and DEFrI-I,HEATFLOW,ANDAGE
Sclater,
1986]. Acrosstheaxial
meanbasementdepthincreases
valley[CAYTROUGH, 1979],Depths
from about3500 m at 40 km west
tobasement, heat flow, andage ofcrust producedata
oftheaxis
toamaximum ofabout4600 matthe present axis
(0 spreadingcenterinthedeep oceans show predictable
relationships
km),then
rises
sharplytoabout3200 mat15kmeast oftheaxis. [Sclater
etal.,1971;
Davis and Lister,1974;Parsons and Sclater,
Beyondthis
pointtotheeast,
basementslopesaway from theaxistheoretical
1977;
Lister,1977]. These relationships
have been matched to
modelsof oceanfloor formationby ParsonsandSclater
to a mean depthof about5000 m at 165 km, with local relief
ranging
from
1500to300 mand increasing
toward the axis.Eastwhich
[1977]using the plate concept presented byMcKenzie [1967],in
heat loss in the vertical direction is the dominant mechanism
of 165kmeastoftheaxis, basement again shallows toa forcooling.
minimummeandepthof about4600 m at a distanceof 200 km,
The major difficulty todetermining marginalbasin
thendeepens
eastwardtoamean depthofabout 5400 mattheage from heat flow and depth isthat many ofthese basinsarevery
eastern
limitofoceanic
crust
atabout
450km.Local
basement
narrow
and/or
verysmall,
anditbecomes
inappropriate
toassume
relief
across
this
segment
rangesfrom
300to700m. that
heatislostinthevertical direction
only.Inthese cases a
Thehighrelief
andappreciable
subsidence
ofcrust
underlying
thesignificant
portion
ofcrustal heatmay belostbylateralconduction
central
trough,
incontrast
tothelesser
relief
andlittle
increase
in aswell.
BoernerandSclater[1987] haveconsidered
thisproblemof ocean
depth with increasingdistancefrom the axis within the outer
floorcreatedby a shortaxisof spreading
isolatedwithinolder,cold
trough,indicatethatspreadingduringthesecond(andpresent)stage
oceanic or continental lithosphere.They have calculated the
of openingwasslowerthanthat duringthe earlierstage. The fact
subsidenceand heat loss along the centerlines of severalshort
that the points at which basementslope and characterchange
coincide with those locationswhere the width of the Cayman
finite lengthaxesof spreadingandhavecomparedthesewith the
Trough changesfrom 70-80 km acrossits outersectionsto 100- subsidenceand heat losspredictedby the standardplate model
[McKenzie, 1967]. Their calculationsshowthat the thermaleffects
110 km acrossits centralsection(Figure3) is furtherevidencethat
of lateral heat losscan be relateddirectlyto basinwidth. Lateral
the changein spreadinginvolveda distinctrearrangement of the
spreading
system.
Thesmall
topographic
risecentered
ataboutheat
loss
atspreading
axes
less
than
200kmlong
has
amarked
and
200kmeast
ofthepresent
axis
couldrepresent
anabandoned
measurable
effect
onboth
crustal
subsidence
and
heat
flow
(Figure
spreading
axis,
whichwould
implyfurther
thatthis
rearrangement
5).
involved a ridge jump. As demonstratedbelow, magnetic For example,20-m.y. old crustformed at a spreadingaxis of
"infinite"length,thestandardplatemodelof McKenzie[1967],lies
anomaliessuggestthat thisis the case.
at a depth of about4200 m (Figure 5). Crust of the same age
formedat an isolatedaxis 100 km long andlosingheatlaterallyas
HEATFLOW well asverticallywouldlie 200 m deeper,andsimilarcrust formed
at an axis only 50 km long wouldbe a kilometerdeeper. At 40
Of thefewpublished heatflow
measurements fromtheCaymanMa, crustformed atanaxis100kmlongwouldlie450m deeper
Trough[Eppetal., 1972;Ericksonetal., 1972](seeTable1),only thanthatformedat a typicalmid-ocean
spreading
axis. Vertical
fivelie overapparently
well-sedimented
crmtandcanbeconsideredheatflow showssimilarrelations withdecreasing
spreadingaxis
free of hydrothermal
effects(Figure4). The qualityof these length.At 40 Ma, heatflow in crustformedat a 100-km-long
measurements
rangesfrompoorto excellent.The onepotentially isolatedaxiswouldbe 10mW m-2 lessthanthatin crustformedat
usefulmeasurement
eastof theaxishasa valueof 67.3mWm-2 amid-oceanic
spreading
center.
2146 ROSEN• ET AD: AGE AND SPREADINGHISTORY OF THE CAYMAN TROUGH

DISTANCE IN KILOMETERS
-600 -500 -400 -300 -200 -100 0 100 200 300 400 500
I I I I I I I I I I I
160 160
I
I
? RIFT AXIS

E
:120- 120

89.0 94.4 94
•: 6o-
o
56.0
72.7
•.08•
29.149.3
9•8 67.3 80
I-- 40- 40

0 • • • 0

-z
,,CI -lENT
,
GE R•

SWANISLANDSFRACTURE
ZO! C-•4
BI NICARAGUA RISE

Fig. 4. Sketchmapandprofileof published


heatflowmeasurements
[Eppet al., 1972;Ericksonet al., 1972]. Well-sedimented
areasof thetroughareshaded.

Becausethe CaymanTroughhasthroughoutits historyconsisted to the west was betweenabout5.5 mm/yr (110 km axis) and 7
of a 100 km or less spreadingaxis terminatingagainstlong mm/yr(80 km axis)betweenthepresentaxisanda point220 km
transformfaults,with bothspreadingaxisandtransforms isolated west. Beyond270 km the half ratecannotbe determinedfor lack
within old, cold lithosphere,it is probablethat lateral heat loss of measureddepths.Measureddepthsbetween220 and270 km are
plays a significantrole in the thermalevolutionof its lithosphere.
This is initially evident in the fact that although the trough
displays all the major characteristicsof an oceanic spreading DEPTH

system,includingcrustalthicknesses [Ewinget al., 1960], crustal


depthsaretoo deepfor theirapparentages[Holcombeet al., 1973;
CAYTROUGH, 1979]. Becausethe width of oceaniccrustbeneath
the CaymanTrough is well defined,100-110 km acrossits central
part and70-80 km acrossits outerparts,the effectsof lateralheat
lossareeasilyfactoredinto calculations of its crustalagebasedon 4000 - INFINITE
LENGTH
basement depthandheatflow. \
Crustalagesalongthe CaymanTroughascalculatedfrom mean 4500- \ \.,..
basementdepths(Figure3) andobservedheatflow (Figure4) are
plotted in Figure 6 againstdistancefrom the presentspreading
axis. Ages have beencalculatedfor two axis lengthsor trough
widths,80 and 110 kin, which approximatethe rangeof width of
the trough. Ages calculatedfor an axis 80 km long are younger
thanagescalculatedfor equivalentpointsassuming a 110-km-long
axisbecausea narrowertroughwouldlosemoreheatlaterallyand
consequently wouldsubsidefaster.
Despitesomescatter,thesederivedagesdescribedistinctlinear
trends,which are approximations of spreading rate. Assming a
140
-• HEAT
FLOW
\\\ /
spreadingaxis 80 km long, the distributionof ageseastof the \\k
presentaxisindicatesspreading half ratesof about6 mm/yrout to ,oo_
about 165 km and about 13 mm/yr beyond 165 km. The
E 60-
equivalentrates for an axis 110 km long are about5 and 11
40
mm/yr. The equivalent ratesfor a 110-km-longaxisareabout5.5
and 5 mm/yr, respectively.The one heat flow measurement on 20

well-sedimentedcrust to the east of the axis yields an age 0 i i i [ i 1

0 20 40 60 80 100 120
correlativeto depthages.Basementsubsidence alonepredictsthat
theoldest
crust
attheeastern
endofthetrough
isolder
than
40Ma AGE INMILLION
YEARS
(Eocene)
andthatthechange
introughopening
rateoccurred
at Fig.
5. Theoretical depth
versus age andheat flowversus
ageprofries for
oceaniccrest within small isolatedbasins[from Boerher and Sclater,
about30Ma(Oligocene). 1987].The curves showprofiles forspreading axes
ofseverallengthsall
Thepicture
ofspreading
tothewestoftheaxisislessclear.The opening
atarateof30mm/yr.
Subsidence
andheat
flowarecalculated
trendsof agesderivedfromdepthindicatethattheopeninghalf rate along
linesperpendicular
toaxesmidpoints.
ROSENCRAN1Z
ETAL' AGEANDSPREADING
HISTORY
OFTIlECAYMANTROUGH 2147

70 - - 7o
110 Km AXIS

- 6o
60 -

ß ß - 50
ß

ß
10.9
(+2.0)mm/yr

...
m 40 ß
- 40

z e\ ß
0
- 30
- 30 -

z
- 20 -
-k ßß :.?" - 20

5.5(-+0.5)mm/yr
-'---•'•, ß/• 4.7(-+0.3)mm/yr
10 -
'/: - 10

WEST EAST
I
i i i 0

60 - - 6O
80 Km AXIS

50 - - 5O

m
z
40 -
ß ee 13.0
(-+2.2)mm/yr - 40

O
- 30 - - 30

z
- 20
- 20 -

. 7.0(-+0.5)m'•m/yr
ß ß 6.1(-+0.5)mm/yr
10 - 10
ee •d •
WEST ß EAST
I I I I I I I I I I 0

-500 -400 -300 -200 -100 0 100 200 300 400 500
DISTANCE IN KILOMETERS

Fig.6. Crustal
agesascalculated
fromdepths
tobasement
(dots)
andheat
flow(stars),
plotted
against
distance
from
thespreading
axis( located
at0 km).Theages
intheupper
panel
arecalculated
assuming
anaxislength
of110km:those
inthelower
panel
are
based uponan80-kmaxial
length.
Depth-based
ages
arecalculated
fromthemean depths
showninFigure
3. Calculated
depth-
based
agetrends,
i.e.,spreading
rates,
areshown
assolidlines.

apparently
anomalousandyieldexcessively
oldages;thisdeepalthoughtheregional
magnetic
anomalieshadbeen previously
basementmaydescribeasmall
fracture
zone
ormayreflectcrustalmapped
anddiscussed
byEwingetal.[1960],
Bowin[1968],and
faulting
relatedtostrike-slip
processes
along
theSwan IslandsGoughandHeirtzler
[1969].
Macdonald andHolcombe [1978]
transform
faultsystem
tothesouth.
Thewestem
heat
flowderived analyzed
indetail
whattheyconsideredthebest
profile
ofthose
agesshow a largescatter
withabroad
trend
ofincreasingage collected
bytheUSNS Wilkes
in1973(profile
1inFigures7-9).
westward.
However,
heatflowvalues
aresignificantly
higher
and Theynotedthatthemagnetics
on theremaining
profileswere
derivedages
younger
than might
beexpected
assuming
symmetrical
"clearly
disturbed"
wheretheprofiles
crossed
ashortoffset
ofthe
spreading
about
thepresent
axis. spreading
axis[seeCAYTROUGH, 1979].Onthebasis of
It isimportant
tonotethatbecause
relative
subsidence
isabetteranomalyidentifications
alongthisnorthern
line,theycalculated
indication
ofcrustalagethandepth alone,relative
depth-derived
thatthetrough opened at a rateof about
20mm/yrbetween the
agesgivea betterindication
of overallagethandoindividualPresent and2.4Maandabout 40mm/yr between 2.4and8.4Ma.
calculated
ages.Agesshown in Figure6 arecalculated
onthe However, RosencrantzandSclater [1986]argued,onthebasis
of
assumptionthatzeroagecrust
liesatadepth of2570m,which is basement depth andcrustal
subsidence,thatthetrough
openedata
theequilibrium
axialdepth
calculatedbyboththeMcKenzie[1967]much slower rateduringthepast30m.y.,about 15mrn/yr,
and
andBoerner andSclater
[1987]platecooling
models.Thepresencethatpartof themagnetic anomaly sequencecouldbereasonably
of theaxialvalleyisignored.Axialdepths
actually
varyaboutthis interpreted
toreflectthisslower
rate.
number byseveralhundred meters,andit islikelythattheCayman Marinemagnetic anomalyprofiles
whichtransect theCayman
axisisdeeper.Assuming a deeperaxisreduces agesrelativeto Troughareshown astwosetsofprofilesin Figure7. Profiles
in
depths.In Figure6 thiswouldhavetheeffectof moving allthe Figure
7aincludethosewhichtrendsubparallel
tothelongitudinal
agepointsdownward (toward younger values)butwould notchangeaxisofthetrough;those
whichcross thetrough athighangles are
therelative
distributions
of ages(i.e.,rates). shownin Figure7b. Althoughtheregional magnetic fieldhas
beensubtracted
from thesedata,the profilescollectivelyshowa
MARINE
M•GNET•C
AnOMAtmS slight
remaining
regional
gradient
sloping
tothesouth
andeast.
No attempthasbeenmadeto rectifyprofilecrossings.
Magnetic
Macdonald
andHolcombe
[1978]werethefirsttoexamine
and edgeeffectsproduced
by thesteep,fault-bounded
wallsof the
identify
marine
magnetic
anomalies
withintheCayman
Trough,trough
(alinear
magnetic
highandlowwhich
parallel
thenorthem
2148 ROSENCRAN1Z
ETAL.:AGE
ANDSPREADING
HISTORY
OFTI-IE
CAYMAN
TROUGH

'o'".
o '-.
" • ' I! .....
.. '"-...•
'O'-.
• / i: • .'• "..
....
..".? , o.... ß
....
:...;
o
,- •

! •r .:I :

'
I
'"..
l ii.0
:1

il .

I :• ' b.I

! !
! ! :

' I

0 I ' ".
• • I 0 '.

". Zl

I :
I ß
I
I :

i '1
I-J --
o ."
i •-

, i¸

• ..
I (
'/U 009 v.. ,' .
: I
: I
i I ' --
:' !• ::

7 .......
7
ROSENCRANIZ ET AL.' AGE AND SPREADING HISTORY OF THE CAYMAN TROUGH 2149

and southernwalls, respectively)do not appear to perturb the


regionalfield beyond35 km from the walls and so do not affect
thoseprofileswhich trendlongitudinalto the trough[Rosset al.,
1986].
Magnetic anomaliesalong profiles parallel to the Cayman
Trough (Figure 7a) show poor definition,primarily becausethe
anomaliesare predominantlyof low amplitude. As a whole, the
anomaliesshowno obvioussymmetricalpatternacrossthepresent
spreadingaxis. However,the setof magneticprofilesdo describe
recognizable lineations.The spreadingaxis,identifiedandlocated
by bathymeticsurveysandnear-bottomobservations [Holcombeet
al., 1973; CAYTROUGH, 1979; Stroupand Fox, 1981], showsa
distinctlinearcentralanomaly,idemifiedaslineationA in Figures
7a and 7c. West of the axis, the magneticfield is dominatedby
two prominentlineations,identifiedasB andC, whichstrikeN-S,
parallelto the presentaxis. Theseextendacrossall the profiles.
The field also shows a number of lesser but still distinct N-S
lineationswhich extendpartway acrossthe profiles (Figure 7c).
The field east of the axis shows little in the way of strong
lineationexceptat D, locatedimmediatelyeastof the axis. Eastof
D, the two southernprofiles show four distinct lineations,but
thesehave no apparentcorrelationto anomaliesalong the two
northernprofries.
Nevertheless,the anomaly pattern containsfeatureswhich
suggestthat its origin lies with seafloorspreading.LineationA
clearly correspondsspatially to the presentextrusionaxis and
temporallyto a centralanomaly.The senseof offsetalongthis
lineationis the same as that mappedfor the spreadingaxis
underneath [CAYTROUGH, 1979]. All lineationslie subparallel
to thepresentspreadingaxisandstrikesubperpendicular
to trough
openingdirectionasindicatedby the orientationsof the transform
faultsboundingthe trough. Despitetheir predominantly
low
amplitude,the singleanomalyprofileslocatedat the easternand
westernendsof the trough(E-E' to the westandF-F' to the eastin
Figure7a) show patterns,
suggesting
thatthesetwosequences
are
symmetrical.
The differencein magneticlineationpatternbetweenthenorthern
andsouthern setsof profilessuggests
thatthetwo arepartiallyand
locallyoffsetalonglineslocatedmedialto andorientedparallelto
the trough. This offset would lie alongthe projectionof the
right-lateraldisplacementof the presentextrusionaxis. This
suggeststhat spreading has intermittently included axial
discontinuitiesor smalltransformzones.However,awayfromthe
offsetsin themagneticpatternhaveno obviousexpression in the
underlyingbathymetry.

MAGNETIC ANOMALY IDENTI•CATIONS

Five of the longitudinalmagneticprofilesshownin Figures7a


and 7c are comparedwith series of calculated, or synthetic,
magneticprofiles (Figure 8). The observedprofiles are those
collected by the USNS Wilkes in 1973 [see Holcombe et al.,
1973]. The syntheticprofilesare calculatedby the fast Fourier
transform technique of SchoutenandMcCamy[1972]appliedto the
reversalstime scalepresentedby Berggrenet al. [1985]. These
calculationsassumecrustalmagnetizationparallelto a geocentric
axialdipolefield direction.The calculatedprofilesare adjustedfor
magneticlatitudeandspreadingorientationrelativeto themagnetic
•,u O0 I.
poleby the applicationof appropriatephaseshift. Magneticcrust
is modeled as a 250-m-thick layer of uniform magnetization
[CAYTROUGH, 1979; Macdonald and Holcombe, 1978] and
constantthicknessparallelingbasement.Magneticlineationsand
basementtopography are both assumedto be two dimensional.
2150 ROSENCRAN•
ETAL.: AGEANDSPREADING
HISTORY
OFTHECAYMAN
TROUGH

I
AXIS
13.4mm/yr I 13.4mm/yr

2O I
• 6.Dmm/yr
6.4mm/yr 6.Dmm/yr

8 9 20

7 ,
i E C B AD -- i

300

2A I 2

5 5C 6 6B 7 8/9

6.4mm/yr

WEST 7.2mm/yr I 7.2mm/yr


AXiS EAST
, I
400 3•0 260 •fi0 0 •0 2•)0 3•0 4•0
DISTANCE IN KILOMETERS

Fig.8. A comparisonofselected
observedmagnetic
anomaly profiles
andcalculated
(synthetic)
anomaly
sequences.
Theobserved
profiles
(collected
bytheUSNS Wilkes;seeHolcombe etal.[1973])
areprojected
ontoavertical
plane
striking
080ø,parallel
to
thelength of thetrough.SeeFigure 7 forprofile
locations.Thepresent
spreadingaxisislocated
at0 km. Thesynthetic
sequencesarecalculated
withthefastFourier
method described
bySchouten
andMcCamy [1972]andhavebeensmoothed
bythe
application
of a 2.5-kmGaussian
filter.Thespreading
ratesshown
arethose
used
tofit calculated
toobserved
anomalies.

Thesynthetic
profile
iscalculated
at1-kmsample
intervals,
whichcalculated
sequences
withtheassumption
ofasynunetric
spreading
approximates
the1- to 2-kmsample
intervalof theobservedacross
thepresent
axis. However,
isolated
portionsof the
profries. observed
sequences
doshowacceptable
correlation
tocalculated
Observed
anomalies
wereidentifiedonthebasisof bestvisual sequences.The bestof thesecorrelations,
whichincludea
matches
tocalculated
anomaly sequences.
Theobservedsequencesrecognizable
matchacrossall of theprofiles,
include thecentral
werecomparedto series
of synthetic
profiles
calculated
withhalf anomaly(lineation
A) andflankinganomalies 2A (lineation
D to
spreading
rates
rangingfrom20 mm/yr [MacdonaldandHolcombe, theeast),andtheanomaly sequence 6-6A-6Btothewestof the
1978]to5 mm/yr,
theapproximate
lower
limitofratepredicted
by axis(Figure8). Anomaly
6 corresponds
to lineation
C. On
crustalsubsidence.
Synthetic
sequences
werealsovariedwith profiles
1 and2 tothewestof theaxis,anomalies
5 to5Carealso
respect
towidthofthezoneofcrustal
accretion
bytheapplication
recognizable,
asareanomalies
3A,4,4A,and5 onprofile
5. This
ofGaussian
filters
withhalfwidths
ranging
from0 to8 km. suggests
thatlineation
B corresponds
toanomaly5.
Thefullsetof observed
anomalies
westtoeastacross
thetrough To theeastof theaxis,correlations
between
theobserved
and
showsnoacceptable
matchtoanylinear,
continuous
calculated
calculated
sequences
asshown inFigure
8 areslight.
Theprofiles
anomalysequence
centered
onandreflected
across
thepresentshow
nowell-developed
lineations,
exceptforlineation
D andfor
extrustion
axis.Nordoobserved
sequences
reasonably
matchthepartial
lineation
extending
across
profiles
2,4, and5 at110km
ROSENCRAN1Z
ET AL.: AGEAND SPREADINGHISTORYOFTHECAYMAN TROUGH 2151

55

//
/
/ 50
/
/
/
/
/

,•, //
,• /
rr // 45

rn ////
o // ,,• ///
(/) //
//// / 3.1 ////
/ //
40

/// .j• .///


iI /////• .////// 35

/ // /
30 C•
; ; /
iI i/ // // z
/i ///
i i //
;; •/
i II / / 25 r-
0
;/ ? z

II i //
iI ,//i / / 20 :=
II /I ß
II /I/

iI///
15

//
Iß 10
;/
o

WEST - 6.0 6.0 EAST


PLEISTOCENE

-500 -400 -300 -200 - 100 100 200 300 400 500

DISTANCE IN KILOMETERS

Fig. 9. Agesof pickedanomalies plottedasa functionof distancefrom thepresentspreading axis. Solid andopensymbols
identifylineationsandisolatedanomalies, respectively.
Spreading
rates(in millimeters
peryear)arecalculatedasbestfit trendsand
drawnassolidlines. The dashed linesdescribethespreadingratescalculated
fromcrustalsubsidence (Figure6). Thetimescaleat
theleft is drawnfromBerggrenet al. [ 1985].

eastof the axis. The broadmagnetichighscenteredat 75 km on This distributionof identifiablelineationsshowsa numberof


profiles4 and5 areapparently theeasternequivalentsof thesimilar interesting features. First, the overall pattern of observed
highslocatedto the westof the axis, whichlie betweenanomalieslineations consistsof separateidentifiable linear sequences,
5 and 6. On this basis,the lineationat 110 km is tentatively separated by lineationswhichdo not correlatewith the adjacent
identifiedas anomaly6 and the adjacentanomalyto the eastas linearsequences. Second,spreading ratesderivedfromthebestfits
anomaly 6B. of theseobserved andsyntheticlinearsequences areaboutthesame
Withintheouterpartsof thetrough,thesequences of anomaliesfor equivalentsequences on eithersideof thepresentaxis. These
between300 and 475 km west of the axis (betweenE andE' in rates are also compatible with those suggestedby crustal
Figures7a,7c, and8) andbetween275 and470 km eastof theaxis subsidence. Third,theidentifiablesequences aresymmetrical to
(betweenF andF')apparentlycorrelateto thecalculated sequencecounterparts on the oppositeside of the presentaxis, but their
13-19, althoughthismustbe considered tenuousin view of the centersof symmetry(reflection)are offsetrelativebothto other
attenuated amplitudesof the observedsequences.However,the sequence pairsandto the presentextrusiveaxis. Thesefeatures
observationthat the large anomalylocatedabout220 km eastof indicatethat the overall processof crustalaccretionwithin the
the axis could representanomalies11 and 12 and that the troughhasinvolvedseveralridgejumps.
easternmost anomalyapparently corresponds to anomaly20 lends Introducingridge jumps into the particular distributionof
supportto thisinterpretation. anomalies acrossthe Cayman Trough yields an essentially
Two anomalies withinthe sequence arethemselves apparently completemagneticanomalysequence from a centralanomalyto
anomalousand not related to seafloorspreading. These are anomaly 19, plus anomaly 20 to the east. This is shown in
hatchuredin Figure8. Both showsignificantlygreateramplitudes Figure9, whereinanomaliesareplottedagainstdistancefrom the
tt,an the anomalieselsewherein the trough. The easternanomaly presentspreadingaxis,andin Figure 10, whichis a reconstruction
maybe relatedto a seamount(?) locatedat thenorthernedgeof the of tl;emagneticanomalyprofile.The centralanomalycorrelatesto
trough(Figure3). The westernanomalyis clearlycircularin plan the presentextrusionaxis as mappedby CAYTROUGH [1979].
butcannotberelatedto anydistinctunderlying bathymetricfeature. The anomaliesflankingthe centralanomalycorrelateto anomaly
2152 ROSENCRAN•
ETAL.: AG• ANDSPREADING
HISTORY
OFTHECAYMANTROUGH
ROSENCRAN• ET AD: AGE AND SPREADING HISTORY OF TttE CAYMAN TROUGH 2153

2A. Anomalies3A to 6B to thewestof thepresentaxisbetween Spreadingcontinued


atthisnewaxisuntilabout26 Ma (Chattiaa),
60 and200km describea stronglineartrendwitha spreading
half whentheaxisagainshifted,thistimeabout30 km to theeast.
rate (slope)of 7.7 mm/yr (Figure9). A conjugate trendis At26 Ma (Chattian),
spreadingin theCayman Troughundergoes
describedto theeastof thepresent
axisbetween
20 and160kmby a majorchange.Therateof openingslowsto between15 and16
anomalies6 and6B plustheanomalies tentatively
identifiedas5, mm/yr(Figure9). The spreading axislengthens
to thesouth,and
5A, and7. This trendyieldsa spreading
half rateof 7.9 mrn/yr. the widthof thetroughincreases
from about80 km to about110
Thedownward (towardyounger ages)extrapolation
of thistrend km (Figure11). Thisincrease of axislengthapparently
is either
suggeststhatanomaly4A lies at thesamedistance
fromthe axis in responseto, or initiates,a southward jumpof theSwanIslands
as the anomalyidentifiedas 2A, which would indicatethat transform faultfroma positionnorthof thepresentSwanIslands
anomaly4A occupiedthe siteof thelatestridgejump. Anomaly to a positionsouthof the islands.This jump transfersthe Swan
4A andpartof anomaly5 havebeenoverprinted by the sequenceIslands(tectonic)Block (SIB) from the NicaraguaRise andthe
2A-centralanomaly-2A.Thisextrapolation furthersuggests
that Caribbean plateto theNorthAmericanplate. Axispropagation,
thisridgejump occurredat anomaly3A, thata fossilridgeis transform faultrepositioning,
andslowing
of spreading
all implya
locatedat anomaly3A at 50 km westof thepresent
axis,andthat significantadjustmentof North American-Caribbeanrelative
the shortlineariondescribedby the small anomalyimmediately motionat this time.
westof the westernanomaly2A on profiles1-4 (Figure8) Veryslowopening
of thetroughcontinued
untilanomaly
3A
represents
anomaly
4. (5.5 Ma, Messinian),
whenthespreading
axisshiftedto theeast
Although
it is notknownwhether
theanomalies
seenalongon about35-40km. Theirregularities
in anomalies
alongprofiles
profile2 between200-450km westof and160450 km eastof the crossing thecentraltrough(Figures7 and8) suggest thatspreading
presentaxisdescribemagneticlineations,thesebestrepresent the duringthistimemayhavebeenneithercontinuous nor orthogonal
anomalysequence 13-19,plusanomaly20 to theeast.Identifiedas to the openingdirection;this is reflectedin the interpretation
of
such,theseshowlinear,symmetrical trendsatspreading halframs anomalydistribution shownin Figure11. This irregularityof
of 13.1mm/yr(Figure9). The pairof high-amplitude anomalies spreading may continueat present,asis indicatedby theskewed
locatedbetween200 and240km arebestinterpreted asanomaly shapeandoffsetof thecentralanomaly (Figure8) [Macdonald and
11-12andits conjugate12-11,reflectedabouta fossilspreadingHolcornbe,1978; CAYTROUGH, 1979]. The rate of trough
centerlocatedat approximately 220km. Thisis supported by the openingmayhaveslowedfurther,to about12 mm/yr,sincethe
fact that the location of this pair corresponds to the small latestridgejump,butthiscannotbeconfirmed withpresent data.
topographic highseenon thebathymetric profiles(Figure3). To
the west,the sequenceof low-amplitudeanomalies between220 DISCUSSION
and300kmcannot
relativeto adjacent
beidentified
sequences
directly,buttheirpositions
indicatethattheserepresent
anomalies
Theoverall definition
ofmarine magnetic anomalies inthe
8and 9reflectedacrossafossilaxislocatedatapproximately 260observed
CaymanTrough isnotnearly asgood asthatofanomalies
elsewherein deep oceanbasinsat equivalentlatitudes.
kmwest oftheaxis (Figure9).Part ofthis spreadingsequence
may have been isolatedto the eastof the presentaxis by the
Whether
this lack
ofdefinitionissuch astopreclude reasonable
interpretationremains a matter of question. Macdonald and
subsequentridge jump (Figures 9 and 10), although this
Holcornbe[ 1978] suggestedthat much of the sequencehas been
interpretationis basedon the recognitionof apparentcrustal
"excess"between anomalies 7 and 11/12 to the cast of the axis and
disturbedby anomalousspreadingprocesses, trar,sformprocesses,
not on the identification of anomalies.
or off-axis volcanism. We agreethat at leasttwo major anomalies
are probablynot relatedto axial magmaticprocess(seeFigures8
and 10). We also recognize that the atypical crustal structure
PROPOSEDOPENING HISTORY OF THE CAYMAN TROUGH
describedby CAYTROUGH [1979] and Stroupand Fox [1981]
may yield an atypicalmagneticsignature. However, the fact that
observedsequenceslocally show goodcon'elationwith synthetic
An openinghistory of the CaymanTrough is presentedas a sequences calculated
fromstandard crustalmagneticmodelsplusthe
seriesof simple paleogeographicreconstructions
in Figure 11. fact that a fair reproductionof the observedsequencecan be
Thesereconstructions are basedupon our magneticanomaly generated with a simplecrustalmodelincorporating theproposed
identificationsas shownin Figures9 and 10 plus the overall ridgejumps(Figure10)suggestthatCaymanTroughmagnetics
mappedsmacture of thetroughasshownin Figure2. may be typicalfor very slowspreading at low magneticlatitude
The time of openingof the CaymanTrough is not directly andthatourinterpretation is correct.
constrainedby availablemagneticdata,although thepresence of Theargument forveryslowspreading issupported by theoverall
anomaly20 (45.5 Ma, Lutetian)at theeasternendof the trough correlation of observed depthsto basement withbasement depths
(Figure10)wouldrequirethatopening occurredpriortolatemiddle curvescalculated frommagneticage(Figure12). The matchof
Eocene.Rosencrantz andSclater[1986]arguethatthewestern and observed to theoreticaldepthis especially goodacross theouter
easternmarginsof the CaymanTroughlie at distances of about trough,bothto theeastandwest.Thecalculated subsidence
curves
500 and460 km, respectively, fromthepresentaxis. Ope-ningreproduce thetopographic highleftby thefossilspreading centerat
ratesas derivedfrom the anomalytrendsandprojectedto the 220l:meastof theaxis. Observed basement depthstothewestof
marginsindicatethatthetroughErrstopenedaboutor priorto 45- the presentaxis are deeperthanpredictedby the magnetics,
50 Ma (Figure9). especiallyacross theintervalbetween220 and300km westof the
Betweenanomaly19 (44 Ma, lateLutetiaa)andthefirstridge axis.It isclear,however, thattheasym_metry of subsidenceacross
jumpat 31 Ma (lateRupelian) thetroughapparentlyopenedat a thepresentaxisis duemainlyto thelatesteastwardjumpof the
rateof between 25 and30 mm/yr.Thewidthof thetroughdinringaxis.The overallmismatch of depthsacross thecentraltrough
thisperiodwasabout80 km. TheSwanIslandstransform fault suggeststhatthe lengthof the spreading axismayin factbe
lay to the northof the presentSwanIslands. At 31 Ma (late shorter
thanmodeled orthatlateralheatlosses
maybegreater than
Rupelian)the spreadingaxis shiftedwestwardabout30 km. predicted bytheBoernerandSclater[1987]coolingmodel.
2154 ROSEN• ETAL.'AGEANDSPREADING
HISTORY
OFTHECAYMAN
TROUGH

B
• /• • • ORIENTE
TiANSFORM
FAULT
/
PRESENTß----m SWANISLANDS
TRANSFORM
FAULT
ß 2AI 2A

A3a (5.5 MA) RidgeJump

18
19
20

A5 (9.5 MA)

Bc• •/,•• / 19
18 15
13119
818
9l•-•

A6(20MA) •

19 18 1513

A7-8(26MA) Ridge
Jump
and
propagation

B //• 49 48 15 4311441
14 43 15
•/<• • . ß ß

A10-11 (31MA) Ridge


Jump

19 20 ?

A19
(44MA) "' •,- >z/•/
Fig.11.Simple
cartoonoutlining
theopening
history
oftheCaymanTrough
asindicated
bymagnetic
anomaly
identifications.
Magneticanomalies
areshown
assolid
rectangles
andsquares.
The bold
solid
line
traces
thepositions
oftheplate
boundary.
Fossil
ridge
axes
areshown
asdashed
lines.
Rift-related
faults
areshown
ashatchured
lines,
with
tics
onthedownthrown
block.
Abbreviations
areasfollows:
B,Belize;
C,Cuba;
J,Jamaica;
SIB,Swan
Islands
tectonic
block.
Seetextforfurther
description.

Crustal
ages
determined
fromheat
flowmeasurements
neither
However,
heatflowdata
aresparse
andareunconstrained
asto
fullysupport
norcontradict
thosecalculated
fromcrustal
position
relative
to underlying
basementtopography.
subsidence
and
magnetics.
Heat
flowfromasingle
station
located
Consequently,
thevalues
mayhave
been
perturbed
byhydrothermal
east
oftheaxisoverwell-sedimented
crust
yields
a crustal
age circulation
induced
byunderlying
basement
topography.
We
co•istent
withthatcalculated
fromdepth
andmagnetics,
butthreesuggest
thatthequestion
ofCayman
Trough
heatflowbeshelved
offourapparently
reliable
measurements
west
oftheaxis
yielduntil
newand
better
measurements
areobtained.
ages
younger
than
predicted
bythemagnetics.
Thisintroduces
a Theorigin
oftheridge
jumps
isunknown.
Onepossible
problem
intotheoverall
consistency
ofourinterpretation.
explanation
isthat
the
spreading
center
isopening
soslowly
that
it
ROSENCRAN'rZET AL.- AGE AND SPREADINGHISTORY OF THE CAYMAN TROUGH 2155

cannotmaintaina permanentmagmachamber. Our spreading rate


of 15 rnm/yr lies near thoserates below which mid-oceanic
spreadingtheoreticallycannotoccurbecausethe axial magma
chambermust instantlyfreeze. Sleep[1975] suggeststhat this
limit is 18 mm/yr; Kusznirand Bott [1976] and Kusznir [1980]
o o o o propose10 and 13 mm/yr, respectively,spreadinglimits. In this
o o o o g o
o o oo o
o o o
o
casethe ridgejumpswould represent"jumpstarts"of the dead
, , , , i , , • ,
accretionsystem. In the caseof the CaymanSpreadingCenter,
however, the systemappearsto spreadat very slowrateswithout
interruptionfor about20 m.y. (25-5 Ma), so freezing the axis
would appearnot to be the controlling factor for the Cayman
jumps. Becausethe two earlyjumps at 31 and 27 Ma coincide
with the overallslowdownof openingandwith a wideningof the
trough (lengtheningof the spreadingaxis), the Caymanjumps
probablyreflect repositioningof the North American-Caribbean
plateboundaryandreorganization of relativemotionbetweenthe
two plates.
The historyof CaymanTroughopeningasindicatedby magnetic
anomaliesand crustal subsidenceraises a number of questions
about several current models of the evolution of the northern
Caribbean. For instance,Pindell and Dewey [ 1982] and Sykeset
al. [1982] describethe northern Caribbeanplate boundary as
developing in two distinct stages. The first involved Late
Cretaceous-Paleocenenorthward convergence, collision, and
suturingof the Antillean arc-Caribbeanplate with the Bahamas
platform. This stage ended during latest early Eocene-earliest
middle Eocene time. A period of no apparentplate motion
followed,lastinguntil the latestEocene-earliest
Oligocenetime,
when the Caribbeanplate startedto move eastwardrelativethe the
North and South Americanplates.The westernhalf of this new
strike-slipplate boundaryevolvedas a largepull-apartbasin,the
CaymanTrough. The easternhalf of theboundarydeveloped into a
complexstrike-slipzonedominatedby transpressive structures.
The 45-50 Ma time of Caymantroughopeningsuggested
by
Troughmagneticsis significantlyearlier than the 36-38 Ma start
of eastward platemotionproposed by PindellandDewey[1982]and
Sykeset al. ••982]. It is importantto note that if the Cayman
Troughfurstdevelopedas a pull-apartbasinalongthisnew strike-
slipboundary,thentimeof openingdoesnotnecessarily constrain
the start of eastwardrelative motion, which may have occurred
earlier.
Recentteevaluationsof stratigraphicage in Cuba indicatethat
convergentmovementin westernand centralCubacontinueduntil
middleEocene[Pszczolkowski andFlores,1986]. Consequently,
the later stageof easternrelativemotionof the Caribbeanplate
would appearto have startedeither before or at the sametime as
o
theearlierstageof convergence
ended.Thiswouldsuggest thatthe
two stages are not separate events but describe parts of a
continuum of motion of the Caribbean relative to North America
(seeMalfai• andDinkiernan[1972],for instance).
_j oø If the CaymanSpreadingCenterridge jumpsmark timesof
0 o
Caribbeanplate andplateboundaryreorganization, thenthe two
z
jumps during the Oligocene record an event which is not
0 0 0 0 0
specificallyrecognizedelsewherealongthe northernCaribbean
0 0 0 0 0 0 0 0 0 0
0 •) 0 •) 0 •) 0 •) 0 •) boundary.The Neogenehistoryof theeasternhalf of theboundary
• 3-11:tOJ::ld
needs to be reexamined to determine whether or not there was an
$B313•4 NI Hid30
01_L3N•D¾•
important reorganizationat that time. Similarly, the jump at
about5 Ma andtheapparent westwardoffsetof thecentralanomaly
suggestthat the boundaryhas eitherjust reorganizedor is in the
processof reorganizing.The argumentsput forthby Sykeset al.
[1982] to explainthe Plioceneslowdownof Caymanspreading
proposed by MacdonaldandHolcombe[1978]areconsistent in part
with our interpretationof the Caymanspreadinghistory. This
2156 ROSEN• ET AL.: AGE AND SPREADINGHISTORY OF THE CAYMAN TROUGH

wouldalsomeanthatpastCaribbean-NorthAmerican
displacement
processes,
we suggestthat new surveysof CaymanTrough
ratesarenotnecessarilyreliable
indicators
ofpresentrate. geologyand geophysics
wouldbe appropriate.A thorough
magneticssurvey(preferablyan aeromagnetic survey),including
CONCLUSIONS therift margins,wouldnot onlydescribethe full spreadinghistory
of the troughbut wouldalsobetterdefinethe accretionmechanics
CaymanTroughcrustalagesbasedbothuponcalculations of of very slow spreadingsysterns.A seriesof drill holesacrossthe
crustalsubsidence anduponidentifications of marinemagnetic lengthof thebasinandits eastandwestmarginswouldserveto
anomalysequences indicatethatthetroughhasopenedat slowto constrainboth the age of oceanic crust and the timing and
veryslowratesin twomainstages.The overalldistribution of geometryof the rifting process. Seriesof precisionheat flow
basement depthacrossthetroughis typicalof thatobservedacross measurements within the basin would provide additional
spreadingcentersin deepoceanbasinsbut showsa marked constraintson crustalageandwouldalsotestthe effectsof lateral
asymmetry of subsidence,whereincrustto theeastof thepresent heat lossin small,isolatedoceanbasinsas predictedby Boerner
axishassubsided fartherthanat equivalentdistancesto thewestof and Sclater [1987].
the axis. Basementalsoshowsdistinctbreaksin slopebothto the
east and west of the axis. The distributionof agesbasedupon Acknowledgments. The authorsthank Ruth Buskirk, Larry
subsidence curves which accommodate lateral as well as vertical
Lawyer, and Peter Vogt for their commentson style and content.
coolingof the accretingslabsuggestthatthe troughopenedat a Thesehave significantlyimprovedthe manuscript. This work
rate of about 15 mm/yr during the later and presentstageof was supportedby the Shell Foundationand the University of
opening
andatbetween
25and
30mm/yr
during
theearly
stage.
Texas.
University
ofTexas
Institute
forGeophysics
contribution
Thechange
inrateoccurred
atabout
26Ma(Oligocene). 731.
The pattern of magnetic anomalieswithin the trough is
nonsymmetric relativeto the presentaxis. However,anomalies References
do describedistinct lineationstrendingN-S acrossthe trough,
parallel to the presentaxis. Isolatedsequences of lineationson Berggren,W. A., D. V. Kent, J. J. Flynn, and J. A. van Couvering,
Cenozoicgeochronology, Bull. Geol.Soc.Am., 96, 1407-1418,1985.
both sidesof the axis showgoodcorrelationto calculatedlinear Bizon, G. et J. J., T. Calmus, C. Muller, and B. van den Berghe,
sequences, anda full sequenceof anomaliesfrom a centralanomaly Stratigraphie du Tertiaire du sud d'Hispaniola (Grandes Antilles):
to anomaly 19 can be identified acrossthe trough with the Influencede la tectoniqued6crochantesurla pa16og6ographie
et 11fistoire
introductionof threeridgejumps into the spreadingsequence. s6dimentaire,in Gdodynarniquedes Caraibes, pp. 371-380, Editions
Lineations mapped within the central part of the trough show Technip,Paris,1985.
Boernet, S., and J. G. Sclater, The two-dimensional infinite dyke
reasonablematchesto anomaliesout to 6B, particularlyto thewest problem: An approximatesolutionfor the threedimensionalheatloss
of the present axis. The identification of anomaliesbetween nearthe centerof a shortoceanicspreading
center,in Marine Heat Flow,
anomalies8 and 19 in the outerpartsof the troughis lesscertain, edkedby J. A. Wright andK. E. Louden,CRC Press,BocaRaton,Fla.,
paxZly
because
thereareinsufficient
datatomaplineations. 1987.
inpress
Theoverall
rateofopening
indicated
bythemagnetic
lineations
Bowin, C.O.,Geophysical
Res. 73, 5159-5173, 1968.
study
oftheCayman
Trough,
J.Geophys.
within
thecentral
trough
isabout
15mm/yr.Thisclosely
agreesBurkan,
B.,Neogene
North
America-Caribbean
plate
boundary
across
with the rate calcula•d from subsidence.The rate of opening northernCentralAmerica: Offsetalongthe Polochicfault,
suggested
by anomalieswithintheoutertroughis between25 and Tectonophysics,
99,251-270,1983.
30 mm/yr. Thisalsois consistent
wifl•theapproximate
rate Burke,
K.,C.Cooper,
J.F.Dewey,
P.Mann,
andJ.L.Pindell,
Caribbean
tectonicsand relative plate motions,The Caribbean-SouthAmerican
suggestedby subsidence. The changefrom faster to slower Plate BoundaryandRegionalTectonics,editedby W. E. Bonini, R.B.
spreadingoccurredat about 26 Ma (Oligocene) These rates, Hargravesand R. Shagam,Mem. 162, 31-64, Geol. Soc. Am., Mem.
extrapolatedto the easternand westernmarginsof the trough, 162, 1984.
indicatethatit firstopenedbet,•'een45 and50 Ma (MiddleEocene). CAYTROUGH, Geological and geophysical investigation of the
If our outline of the openinghistory of the Cayman trough,as MidCaymanRiseSpreading Center:Initial resultsandobservations,
in
Deep Drilling Resultsin the Atlantic Ocean: Ocean Crust, Maurice
basedupon the identificationof marine magneticsequences,is Ewing Ser., vol. 2, edited by M. Talwani, C. G. Harrison, and D. E.
correct,thenboth oceaniccrustwithin the troughand its overall Hayes,pp. 66-93, AGU, Washington,D.C., 1979.
structurepotentiallyprovidea greatdeal of informationaboutthe Crough, S. T., The correctionfor sedimentloading on the seafloor,J.
evolutionof the Caribbeanas well as significantinsightinto the Geophys.Res., 88, 6449-6454, 1983.
Davis, E. E., and C. R. B. Lister,Fundamentals
of ridgecresttopography,
evolutionof smallrift basins.In termsof Caribbeangeology,the Earth Planet. Sci. Lett., 21,405-413, 1974.
troughopeninghistoryprovidesquantitativemeasuresof relative Deaton,B.C., and B. Burkart,Time of sinestralslip alongthe Polochic
plate motion along the northernCaribbeanplate boundaryzone, fault of Guatemala,Tectonophysics,102, 297-313, 1984.
whichmay be tied directlyto on-landstudieswithin the boundary Dziewonski, A.M., T.-A. Chou, and J. A. Woodhouse,Determinationof
ea_'thquakesourceparametersfrom waveformdatafor studiesof global
zone. It alsoprovidesconstraints on therelativemovementof the andregionalseismicity,J. Geophys.Res.,86, 2825-2852, 1981.
Caribbeanplate as a whole with respectto surrounding plates, Edgar, N. T., J. I. Ewing, and J. Hennion, Seismic refraction and
includingtheimportantbutpoorlyunderstood changein motionof reflection in the Caribbean Sea, Am. Assoc. Pet. Geol. Bull., 55, 833-
the Caribbeanplate from northward(convergentboundaries)to 870, 1971.
eastward (strike-slip boundaries)in the Eocene. Outside the Emery,K. O., andJ. D. Milliman, Shallowwaterlimestonesfrom slopes
of GrandCaymanIsland,J. Geol., 88, 483-488, 1980.
Caribbeanframe of reference,the troughtakeson importancenot Epp, D., P. J. Grim, andM. G. Langseth,Heat flow in the Caribbeanand
only as a prime exampleof a very slow spreadingsystembut also Gulf of Mexico, J. Geophys.Res., 75, 5655-5669, 1972.
as a smallbut maturemodelrift basinwhosedevelopment may be Erdlac,R. J., Jr., andT. H. Anderson,The Chixoy-Polochic fault andits
reconstructed in full. This evolution may includebasinorigin as associatedfracturesin westernGuatemala,Geol. Soc.Am. Bull., 93, 57-
67, 1982.
a pull-apartbasin. Erickson, A. J., C. E. Helsley, and G. Simmons, Heat flow and
In view of thesepotentialcontributionsto resolvingregional continuous seismicprofilesin the CaymanTroughandYucatanBasin,
tectonics and to better understandingrifting and spreading Geol. Soc. Am. Bull., 83, 1241-1260, 1972.
ROSENCRAN• ET AL.: AGEAND SPREADING
HISTORYOFTI-IECAYMANTROUGH 2157

Ewing,J.,J. Antoine,
andM. Ewing,Geophysicalmeasurementsin the Perfit,M. R. and B.C. Heezen,The geologyandevolution of the
westem CaribbeanSeaandin theGulfofMexico,J. Geophys.
Res.,65, Cayman Trench,Geol.Soc.Am.Bull.,89, 1155-1174,
1978.
4087-4126,1960. Pindell,J., and J. F. Dewey,Penno-Triassic
reconstruction
of western
Gough,
D. I., andJ.R. Heirtzler,
Magnetic
anomalies
andtectonics
ofthe Pangeaandthe evolution
of the Gulf of Mexico/Caribbean
area,
CaymanTrough,Geophys. J. R. Astron.Soc.,18,33-49,1969. Tectonics,
1,179-212,1982.
Holcombe,
T. L., P. R. Vogt,J. E. Matthews, and R. R. Murchison,Pszczolkowski,
A., andR. Flores,Fasestect6nicas
delCreficicoy del
Evidence
for sea-floor
spreadingin theCayman Trough,
EarthPlanet. Pale6genoenCubaoccidentaly central,
Bull.Acad.Pol.Sci.,Ser.Sci.
Sci.Lett.,20, 357-371, 1973. Geol.Geogr.,34, 7-14, 1986.
Jordan,T. H., The present-day motionsof the Caribbean plate,J. Rosencrantz,E., andJ. G. Sdater,Depthandagein theCaymanTrough,
Geophys. Res.,80, 4433-4439,1975. EarthPlanet.Sci.Lett.,79, 133-144,1986.
Kusznir,N.J., Thermalevolution
of theoceaniccrust;itsdependence on Ross, M. I., E. Rosencrantz,C. R. Scotese,and J. Sclater,New
spreading rateandeffectoncrustal structure,
Geophys. J. R. Astron. interpretation
of spreadingratesintheCaymanTrough, Geol.Soc.Am.
Soc.,61, 167-181,1980. Abstr.Programs,18, 733, 1986.
Kusznir,N.J., andM. H. Bott,A thermalstudyof theformation
of Schouten,
H., andK. McCamy,Filteringmarinemagnetic
anomalies,
J.
oceaniccrust,Geophys.
J. R.Astron.
Soc.,47, 83-95,1976. Geophys.
Res.,77,7089-7099,1972.
Land,L.S.,Thefateofreef-derived
sediment
onthenorth
Jamaican
island Schwartz,
D. P., L. S. Cluff,andT. W. Donnelly,
Quaternary
faulting
slope,
Mar.Geol., 29,55-71,1979. along
theCaribbean-NorthAmerican plate
boundary
inCentral
America,
Lister,C. R. B., Estimators
for heatflow anddeeprockproperties
based Tectonophysics, 52, 431-445,1979.
uponboundary
layertheory,
Tectonophysics,
41,157-171,1977. Sclater,J. G., R. N. Anderson,and M. L. Bell, Elevationof ridgesand
Macdonald,K. C., andT.L. Holcombe,Inversionof magneticanomalies evolutionof the centraleasternPacific,J. Geophys.Res., 76, 7888-
andsea-floor spreadingin theCaymanTrough,EarthPlanet.Sci.Lett., 7915, 1971.
40, 407-414, 1978. Sleep,N.H., Formationof oceaniccrust:Somethermalconstraints, J.
Mallair, B. T., and M. G. Dinkelman,Circum-Caribbean tectonicand Geophys.Res.,80, 4037-4042,1975.
igneousactivityandtheevolutionof the Caribbean plate,Geol.Soc. Stroup, J. B., and P. J. Fox, Geologic investigationsin the Cayman
Am. Bull., 83, 251-272, 1972. Trough:Evidencefor thin oceaniccrustalongthe Mid-CaymanRise,J.
Mann, P., K. Burke,andT. Matsumoto,Neotectonics of Hispaniola:plate Geol., 89, 395-420, 1981.
motion,sedimentation andseismicity bend,EarthPlanet. Sykes, L. R., W. R. McCann, and A. L. Kafka, Motion of Caribbean
at a restraining
Sci. Lett., 70, 311-324, 1984. plate duringlast 7 million yearsand implicationsfor earlier Cenozoic
Mann,P., M. R. Hempton,D.C. Bradley,andK. Burke,Development of movements, J. Geophys. Res.,87, 10,656-10,676,1982.
pull-apartbasins, J. Geol.,91,529-554,1983. Wadge,G., andIC Burke,NeogeneCaribbeanplaterotationandassociated
McKenzie, D. P., Someremarkson heatflow and gravityanomalies, J. Central American tectonicevolution, Tectonics,2, 633-643, 1983.
Geophys. Res.,72, 6261-6273,1967.
Molnar, P., and L. R. Sykes,Tectonicsof the Caribbeanand Middle
Americaregions fromfocalmechanisms andseismicity,
Geol.Soc.Am. E. Rosencrantz,M. I. Ross,and J. G. Sclater,Institutefor Geophysics,
Bull., 80, 1639-1684, 1969. Universityof Texasat Austin,8701MopacBlvd.,Austin,TX 78759.
Parsons,B., andJ.G. Sclater,An analysisof the variationof oceanfloor
bathymetry
andheatflow with age,J. Geophys.
Res.,82, 803-827,
1977. (ReceivedApril 2, 1987;
Perfit, M. R., Petrologyand geochemistryof mafic rocksfrom the revisedAugust4, 1987;
Cayman
Trench:
Evidence
forspreading,
Geology,
5, 105-110,1977. acceptedOctober5, 1987.)