GEOPHYSICAL
RESEARCH
LETTERS,
VOL. 14, NO. 6, PAGES
595-598, JUNE1987
LATE
TERTIARY
HISTORY
OF HYDROTHERMAL
DEPOSITION
AT •
RISE, 19øS: CORRELATION TO VOLCANO-TECTONIC
EAST PACIFIC
EVENTS
MitchellLyle
Collegeof Oceanography,
OregonStateUniversity
MargaretLeinen
GraduateSchoolof Oceanography,
Universityof RhodeIsland
Robert M. Owen and David K. Rea
Departmentof AtmosphericandOceanicScience,Universityof Michigan
Abstract.Commonlyit is assumed
thattheintensity
of mid-oceanridgehydrothermal
activityshouldcorrelate
with spreadingrate, since high spreadingrates are an
indication
of largesubcrustal
heatsources
neededfor intense
hydrothermal
activity. We havetestedthishypothesis
by
modelingthedeposition
of hydrothermal
precipitates
from
coresfromDeepSeaDrilling ProjectLeg 92, takenon the
hydrothermal
precipitates
formedin plumesabovethepresent
EastPacific Rise and its ancestor,the Mendoza Rise (part of
the old Pacific-FarallonRidge), and wafted to the west by
deepoceancurrents[Leinen,Rea, et al., 1986]. Sediments
at both sitesare typicalof centralgyre sedimentsexceptfor
the largehydrothermally-derived
comrx>nent.Sedimentation
west flank of the East Pacific Rise at 19øS. Although
rates
arelow,ranging
from15m10Gyr
-1to0.1m106yr
-1,
14, and 9 million years ago, of hydrothermalmanganese
andvary primarilydueto depth-dependent
changesin calcite
preservation. There is very little input of terrigenousor
organicmatter,andno evidencefor diageneticremobilization
of manganese,the most redox-sensitiveof the transition
elements[Leinen,Rea,et a1.,1986;Lyle, 1986; Lyle et al.,
deposition
asmuchasa factorof 20 higherthanequivalent
1986].
Holoceneaccumulation. These eposidesdo not correlate
with spreadingrate changesand insteadseemto occur at
timesof major tectonicreorganizations.We proposethat
The hydrothermal component in Sites 597 and 598
derivesfrom the westwartdrift of hydrothermalprecipitates
in a plumefrom theEastPacificRiseaxis. The plumecanbe
tracedover 1000 km to the west of the East Pacific Rise by
spreading
ratesat theEastPacificRiseandits predecessor,
the MendozaRise, havevariedby only 50% in the last 30
millJoriyears,we foundcertainepisodes,
at about25, 18,
ridge jumps and changes of ridge orientation may
substantially
increase
hydrothermal
activityby fracturingthe
oceancrustand providingseawateraccessto deep-seated
anomalous
3Hecontent
indeep
waters
[Lupton
andCraig,
1981] and may be driven by geothermalheatingat the East
Pacific Rise [Stommel,1982]. As the hydrothermalplume
wafts away from the Rise Axis, hydrothermalprecipitates
heat sources.
settle out to sediments.
Introduction
Model of HydrothermalPlumeDeposition
Hydrothermalcoolingof basalticcrustis ubiquitousat
mid-oceanridges and may accountfor 20% of the total
geothermalheatlostby theearth[Williamsandyon Herzen,
1974]. Elemental exchangebetweenseawaterand basalt
during this process is instrumental in controlling the
compositionof seawaterand how it has varied through
geologictime [Berneret al, 1983; Owen and Rea, 1985].
Several attemptsto model this process have made the
first-order assumption that hydrothermal exchange is
proportionalto spreadingrate [Lyle, 1976; Grahamet al.,
1982;Berneret al., 1983;Sleepet al., 1983]. In thispaper
we report the first time-series measurementsof the
accumulation
of hydrothermalprecipitates
andshowthatthe
late Tertiarygeologichistoryof hydrothermal
activityis, in
fact, unrelatedto changesin spreadingrate. Instead we
observelarge changesof hydrothermalactivity which are
morecloselyassociated
with tectonicreorganizations
of the
mid-oceanridges. Presumablytectonicreorganizations
The massaccumulation
rateof hydrothermal
precipitates
should decreaserapidly from the venting area, since the
amount of mass in a distribution of particle sizes is
preferentially
in thelarger,morereadilysedimented
particles.
This is preciselywhatis observedat 19øSon theEastPacific
Rise. Figure 1 shows measured accumulationrates of
manganese(the major element most highly enriched in
hydrothermalprecipitateswith respectto continentalor
basalticdetritus)on a profile between18øSand 20øSand a
modelledmanganeseaccumulationrate profile [Lyle et al,
1986] assuming1) a constantcompositionof hydrothermal
precipitates[Dymond, 1981], 2) a log-normalparticle size
distribution[Lambertet al, 1981]with a mediangrainsizeof
2 micrometers,
3) Stokessettlingof particlesfromtheplume,
and4) a deepcurrentdirectionnormalto therise axis. With
this type of model, the distancea hydrothermalparticle of
diameterD travelsis givenby:
profoundlyaffectfracturedepthandspacing,resultingin
increased access of seawater to hot rock, and a marked
Dist= V I
increase in the chemical exchangebetween basalt and
K D2
seawater.
DeepSeaDrilling Project(DSDP) Site597 (18ø48.4'S,
129ø46.2'W) and Site 598 (19ø00.3'S, 124ø40.6'W) were
(1)
whereV is the meancurrentvelocitynormalto the rise
crest,I is the injectionheightof hydrothermalparticles,and
K is the Stokes constant. The mass of hydrothermal
drilledduringLeg 92 by theR/V GlomarChallengerin part
to determine the geologic history of deposition of
deposition
at somegivendistance
fromtherisecrestis then
Copyright1987by the American
Geophysical
Union.
calculated from the original particle distribution at the
Papernumber7L7141,
hydrothermal
sourceareaandfrom the massof individual
particlesthatwouldsettleat thisdistance.By manipulating
0094- 8276/87/007L-7141
$03.00
the model we can match today's sedimentary flux of
595
596
Lyle et al.: LateTertiaryHydrothermal
Deposition
midwater
AGE (My)
(Rise Jump) 18.5
I
40•
I'0
5
>'
20
25
I
28 5
I
'
'
15
18.,5
Rise is
however, we can eliminate events that could have been
35
causedby changesin currentdirectionor velocity. We used
E
x
current structure near the East Pacific
probably the most poorly understood of the model
parameters
andit probablyhasnotbeenconstant
for thepast
30 million years. By a simple addition to the model,
a ran•ge
of currentvelocities
in themodel,from1 to 5 cm
30
"'
2,5-
sec-• normal to the ridge axis, and only acceptedas
anomaliesthe episodesof manganeseaccumulationwhich
were higherthan could be accountedfor by this velocity
range. An exampleof this is shownin Figure 2, from Site
z
20-
597.
E
o
-J
In essence, using the current velocity range also
eliminatesthe possibilitythat changesin currentdirection
could have producedthesemanganeseaccumulationrate
1,5-
anomalies. The best available information on mid-water flow
:•
n
1200
I
IOOO
800
600
PALEODISTANCE
400
200
(krn)
Fig.1. Observedaccumulationrate of manganeseat DSDP
Site 597 versus distance from the East Pacific Rise Axis as
compared to a modelled manganeseaccumulation rate
assuminga ridge axis hydrothermalsourcefor manganese.
The heavy line marks the reecordfrom Site 597, which is
brokenwherepaleo-distances
from therisecrestoverlapdue
to the Mendoza-EastPacificRise ridgejump at 18.5 million
yearsago [Mammerickxet al., 1980]. The light line marks
the model accumulation rate. Open circles delineate
measured manganese accumulation rates for Holocene
sediments[Lyle et al., 1986]. Manganeseaccumulation
rates
were modelledby assuminga constantcompositionand a
constant log-normal particle size distribution for
hydrothermalmaterial, and by assuminga constantcurrent
velocitynormalto theridge. SeeLyle et al. [1986] for more
details. Accumulation rates of manganeseat Site 597
betweenabout27 and 24 million yearsago andbetween19
and 17 million yearsago are sufficientlylarge to only have
beencausedby anomalously
highhydrothermal
activity.
at thislatitudesuggests
thatthepresentflow regimeis aswe
havemodelledit, approximately
to thewestandnormalto the
rise crest [Reid, 1982]. Minor increasesof hydrothermal
depositionmay haveoccurredin thepastdue to a mismatch
of modelandthe actualsituation--if theflow now is slightly
off theidealizeddirectionandflow in thepastbecamemore
normal to the ridge. Major changesin current direction
would reducethe hydrothermalmanganeseaccumulation,
however,sincetheplume wouldhavetravelledobliquelyto
the rise axis.Distanceto the sourceregionwould thushave
increaseddramatically.
The sedimentation
parameters
in the modelalsohelp to
excludedoubtfulhydrothermalepisodes. The assumption
in
the modelof Stokessettlingof hydrothermal
plumeparticles
neglects aggregation processes that help to increase
depositionratesof fine particulatematerial. However, the
lack of theseaggregationprocesses
in the modelmaximizes
the depositionin the distal sedimentsgreaterthan 100 km
from the rise axis and thusalso helpsto excludedoubtful
hydrothermalepisodes.
PALEODISTANCETO RISE CREST (km)
.oo
•
I
o,o
.,ooIf$
1986;andFigure1]. We havefoundthroughmanipulation
of the modelthatthe manganese
accumulation
rateprofile
severely limits the range of acceptableparticle size
distributions
andcurrentvelocities.The shapeof theprofile
andthepredicted
massflux changedramatically
withchanges
of theseparameters
[Lyle et al., 1986].
Sinceeachof the DSDP sitesmigratesaway from the
ridgeaxisat the spreadinghalf-rateof theridge,onewould
expectto find a similarpatternof decreasein manganese
accumulationrate upcoreto thatfoundin surfacesediments.
High ratesof manganese
accumulation
shouldbe foundnear
the baseof the sedimentcolumn, when the sedimentswere
closeto the rise axis, and low ratesshouldoccurtoday,
whenbothSites597 and598 aremorethan1000km away
from the East Pacific Rise. In general that is what we
observe,asis illustratedby the accumulation
rateprofileof
manganesefor Site 597, shownin Figure 1. There are two
intervals in this core, however,
where observed
accumulation
ratesof manganese
aresignificantly
higherthan
themodelled:between24 and27 millionyearsago(AN25),
andbetween17 and18.5millionyearsago(AN 18) [Lyleet
al, 1986].
Theseanomaliespersisteven when the deficienciesof
the simple model we use are taken into account. The
D
•
I
,oo
I
NI8
o
riseaxisof 1 cmsec
-1,typical
ofdeep
flow[Lyleetal.,
I
20
z
hydrothermal
manganese
at 19øSwitha medianparticlesize
of 2 micrometers
anda meancurrentvelocitynormalto the
.o,o
•5
D
i-
AN 14 (?)
IO
o
0
..........
•"":•:•:•:•:•:•:'"-'"'""--•'""'"""•'•
•::':•::"'::':•:•:•:•:•:•:•:•:•:•:•:•
•:'""--'"""'"'"'"'--"•:ø'•;••••!•:•:•:•::':':'•::'•::':"•'Present
=Post•i•
I
I
•
I
I0
12
14
16
18
20
2
I
I
I
24
26
28
AGE (M•)
Fig. 2. Anomalousmanganeseaccumulationrates at Site
597, arrivedat by dividing the manganeseaccumulationrate
observed in sediments of a certain age by the model
Holocenerate at the equivalentdistancefrom the rise crest
(seetext andLyle et al. [ 1986] for details). A key variablein
the model is the currentvelocity normal to the rise crest,V.
Becausethe currentvelocity hasprobablynot beenconstant
in the past, we have chosen to use a range of current
velocities
between
1and5cmsec-1
inthemodel.
Episodes
of high hydrothermalproductionare delineatedby time
periodswhere the manganeseaccumulationrate cannotbe
matched usingthiscurrentrange.
Lyle et al.: LateTertiaryHydrothermal
Deposition
The lack of manganeseaccumulationrate anomalies
younger than AN18 at Site 597 reflects the lack of age
definitionneededto outlineeventsin thepelagicclaysof the
upper part of Site 597 and also the decreased model
resolutioh
atdistances
greaterthan600kmfromtherisecrest.
For thesereasonswe alsoanalyzedmaterialfrom Site 598 to
examinethetimeperiodbetween16 millionyearsagoand8
million years ago. Two more episodesof anamolous
manganeseaccumulationrate were discoveredat Site 598,
between12and16millionyearsago(AN14) andbetween8
and 11 millionyearsago(AN9). Bothof thesetime periods
hadmanganese
accumulation
ratesabout5 timesand2 times
higherrespectivelythanthehighestmodelrate. Manganese
accumulation
rateanomaliesof thismagnitude,observedfor
all 4 events,
canonlybeexplained
bymajorincreases
in the
productionof hydrothermalmanganese
andnot by changes
in dispersion
of thehydrot!lermal
effluent[Lyleetal., 1986].
This implies that hydrothermal activity and chemical
exchangebetweenbasaltand seawaterat the East Pacific
Rise/Mendoza
Riseaxismusthavebeensignificantly
greater
duringthetime periodsin question.
Discussion and Conclusions
The manganeseaccumulationrate anomaliesdo not
correlateeitherin timingor in magnitude
with changes
in
spreadingrate at the East Pacific Rise (Figure 3). The
WeSt-flank
spreadinghalf-rateat 19øShasvariedby only
about50% between30 millionyearsagoandthepresentand
hasonlyonepeak,at about7 millionyearsago. Duringthe
interval for which we have accumulation rate data from Sites
597 and598, from28.5millionyearsagoto 8 millionyears
ago, the half-spreadingrate apparentlychangedonly from
about
56km106yr
-1toabout
73km
106yr
-1orabout
30%.
-•o
I
'
I
I
Cocos-NazcaMendoza
-EPR
RiftFormation R•dge
Jumps
•
•
60
Galapagos
R-EPR
R•dge
Jumps
i
•
Rote, EPR.
West Flank, 19øS
(km/My)
TectonicEvents
Selk•rk R - EPR
RidgeJump
Hawaiian
Extrusion
Rote
Per Unit T•rne,
I0
(103
krn3
)
0
Stag
'
iWeokModerate
S•.•ngCircumPoclhc
i
•
Volcorec Episodes
24x
AN25
i
30
•1
25
AN18
IOxPulsesLeg
92
AN9
i
i
20
15
i•
I0
(Roho
ofobserved
Mn
•--No
Datai
5
anomaliesandspreading
ratechanges.
There is better coincidence between the timing of
hydrothermaleventsAN9, AN14, AN18, and AN25 and
times when Pacific plates were undergoingteeionic
reorgani•zation.AN25 and AN18 matchthe bestestimates
when the Cocos-NazcaRift was formed (25 million years
ago;Hey, 1977) andwhenthe spreading
centerjumpedwest
from the Mendoza Ridge to the East Pacific Rise (18.5
million years ago; Mammerickx et al., 1980). There is no
known ridge reorganizationthat correlateswith AN 14, but
thetectonichistory.oftheEastPacificRiseduringthistimeis
not well known (Mammerickx et al., 1980). AN9 may
correlatewith theabandonment
of thefossilGalapagos
Ridge
(Mamme,rickxet al., 1980), though the matchin time is not
very good.
The periodsof high manganeseaccumulationrate may
also correlate in time with other indicators of Pacific -wide
tectonicevents,especiallywith episodesof circum-Pacific
arc andmid-platevolcanism(Figure3; Kennettet al., 1977;
Rea and Scheidegger,1979; Gardner et al, 1985). These
episodesof arc volcanismarepresumablydueto changesin
convergence
ratesand/or directionsat subductionzonesand
should therefore be related to either changesin rate of
spreading
on a globalscaleor changes
in plategeometry.In
thelast20 millionyearstherehasbeenan importantvolcanic
episodeat 14-16 million yearsago correlatingwith AN 14,
anotherless importantepisodebetween 11 and 8 million
yearsagothacorrelatesto AN9, andtwo episodes
for which
we haveno availablemanganese
accumulation
datafrom our
cores,between6 and 4 million yearsago, and between 1.5
million years ago and the present(Kennett et al., 1977).
There may alsobe a regionalvolcanicepisodeat around18
million yearsago and correlatingwith AN 18, as shownby
ash layer data and by the timing of Central American
volcanism(Kennettet al., 1977). Sincewe use manganese
accumulation
rateinformationfrom thepresentto determine
anomaliesin thepast,therecannotbe a completecoincidence
between circum-Pacificvolcanismand mid-oceanridge
hydrothermal
activityor muchof thepastshouldaccumulate
manganese
at a slowerratethanthepresent.
that there is no correlation
between
sea-floorspreading
rateandEastPacificRisehydrothermal
activity;rather,hydrothermalactivitycorrelatesmuchbetter
with periodsof ridge jumps (Owen and Rea, 1985) and to
some extent also with periods of greater circum-Pacific
volcanicactivity. We believethattheextentof hydrothermal
activityat mid-oceanridgesis dependent
not only uponthe
total heat available, which shouldbe proportionalto the
spreadingrate,but alsouponwhethertheheatis accessible
to
convecting seawater. The process of creating a new
spreadingridge, a commonoccurrence
in the historyof the
Pacific Ocean,mustinvolve a sublithospheric
heat source
and subsequentcrackingof an entire plate. The greatly
increasednumberanddepthof fracturesandcracksbrought
aboutin this mannerwould explosemuch more basalt(by
aboutthe cubeof the total fracturelength) to hydrothermal
5x
flux
to
background
flux)
0
AGE (My)
Fig.3. Summary
of volcano-tectonic
events
in thePacific
Basinfor thelast30 millionyearS,fromLeinen,Rea,et al.
[ 1986],Lyle et al. [ 1986],Hey [ 1977],Mammerickx
et al.
[1980],Kennettet al [1977],ReaandScheidegger
[1979],
and Gardner et al [1985].
The half-spreading rate
informationis a combinationof observations
from Leg 92
and Rea and Scheidegger
[1979]. Spreadingrateshave
variedby lessthan50% duringthelast30 millionyears,
muchlessthanhydrothermal
manganese
accumulation
rates
have varied. The timing of hydrothermalpulsesalso
correlates
poorlywith knownspreading
rate variations.
Thereis bettercorrelation
betweenhydrothermal
pulsesand
either circum-Pacific volcanism or known tectonic events at
the East Pacific Rise.
arean orderof magnitude
larger thanspreading
ratechanges
and thereis no temporalcon-elationbetweenhydrothermal
We document
3O
20
iLeg
90I
Clearly,manganese
accumulation
rateanomalies
weobserve
,io•••1
I00
Half
Spreading
80
4
597
alteration.
Thus
the first
direct
measurements
of
a
hydrothermal
accumulation
ratetime-series
at theEastPacific
Riseindicatethatthe assumption
of proportionality
between
hydrothermalactivity and spreadingrate is too simplistic.
However, these results do reveal an area for refinement of
basalt-seawater
exchangemodelsbasedon the encouraging
correlation observedbetween exchangerates and plate
reorganizations.
Acknowledgments.We are gratefulto the captainand
the crew of theR/V Glomar Challengeraswell asthe restof
theLeg 92 shipboard
partyfor thehelpin obtainingthecores
598
Lyle et al.: LateTertiaryHydrothermal
Deposition
and for all the discussionsabouthydrothermalactivity. We
were supportedfor thisprojectby NSF grantsOCE-8410146
(Lyle), OCE-8411054 (Leinen),andOCE-8410034 (Owen).
References
Berner, R.A., A.C. Lasaga, and R.M. Garrels. The
carbonate-silicategeochemicalcycle and its effect on
atmosphericcarbon dioxide over the past 100 million
years. Arner.J. Sci., 283. 641-683, 1983.
Dymond,J. Geochemistryof NazcaPlatesurfacesediments:
an evaluation of hydrothermal, biogenic, detrital, and
hydrogenous
sources.in NazcaPlate:CrustalFormation
and Andean Convergence, edited by L.D. Kulm, J.
Dymond, E. J. Dasch, D.M. Hussong,and R. Roderick.
Geol Soc Amer. Memoir 154, 133-174, 1981.
Gardner, J.V., C.S. Nelson, and P.A. Baker. Distribution
andcharacter
of palegreenlaminaein sedimentfromLord
Howe Rise: a probableLate Neogene and Quaternary
tephrostratigraphic
record. Init. Repts.DSDP Leg 90.
1145- 1159 , U.S. GovernmentPrinting O•fice,
WashingtonD.C., 1985.
Graham, D.W., M.L. Bender, D.F. Williams, and L.D.
Keigwin, Jr. Strontium-calcium ratios in Cenozoic
planktonic foraminifera. Geochim. Cosmochim. Acta.
•
1281-1293, 1982.
Hey, R. Tectonicevolutionof the Cocos-Nazcaspreading
center. Geol. Soc. Amer. Bull.• 83, 1404-1420, 1977.
Kennett, J.P., A.R. McBirney, and R.C. Thunnell.
Episodesof Cenozoic volcanismin the Circum-Pacific
region. J. Volcan. Geotherm.Res.. 2. 145-163, 1977.
Lambert, C.E., C. Jehanno, N. Silverberg, J.C.
Brun-Cottan,andR. Chesselet.Log-normaldistributions
of suspendedparticlesin the open ocean. J. Mar. Res..
39. 77-98, 1981.
Leinen, M.L., D.K. Rea, et al. Initial Ret•orts of the Deet•
o
_
Sea Drilling Proiect.Leg 92. 617 pp, U.S. Government
Printing
Office,•Vashington
D.C., 1986.
Lupton, J.E., and H. Craig. A major helium-3 sourceat
Lyle, M. Major elementcompositionof Leg 92 sediments.
Init, Repts.DSDP Leg 92, 355-370, U.S. Government
PrintingOffice, WashingtonD.C., 1986.
Lyle, M. , R.M. Owen, and M. Leinen. History of
hydrothermal
sedimentation
at theEastPacificRise, 19øS.
Init. Repts. DSDP Leg 92. 585-596, U.S. Government
PrintingOffice, WashingtonD.C., 1986.
Mammerickx, J., E. Herron, and L. Dorman. Evidence for
two fossil spreadingridges in the southeastPacific.
Geol. Soc. Amer. Bull.. 91. 263-271, 1980.
Owen, R.M., and D.K. Rea. Sea-floor hydrothermal
activity links climate to tectonics:the Eocene carbon
dioxidegreenhouse.Science.227. 166-169, 1985.
Rea, D.K., and K.F. Scheidegger. Eastern Pacific
spreadingrate fluctuationand its relationto Pacific area
volcanic episodes. J. Volcan. Geotherm. Res,., 5,
135-148, 1979.
Reid, J.L. Evidence of an effect of heat flux from the East
PacificRiseuponthecharacteristics
of mid-depthwaters.
Geot•hvs. Res. Lett..9. 381-384, 1982.
Sleep,•T.H.,J.L. Morton,L.E. Burns,andT.J. Wolery.
Geophysicalconstraints
on the volumeof hydrothermal
flow at ridgeaxes.in HydrothermalProcesses
at Seafloor
Spreading
Centers
editedby P.A.Rona,K. Bostrom,
L.
Laubier, and K.L. Smith, Jr., Plenum Press,New York,
53-70,1983.
Storereel, H. Is the south Pacific helium-3 plume
dynamically active?Earth Planet. Sci. Lett.. 61.63-67,
1982.
Williams, D.L., and R.P. von Herzen. Heat loss from the
earth: New estimate. Geology.2. 327-328,1974.
Margaret Leinen, Graduate School of Oceanography,
Universityof RhodeIsland,Kingston,RI 02881.
Mitchell Lyle, College of Oceanography,Oregon State
University, Corvallis, OR 97331.
RobertOwen andDavid Rea, Departmentof Atmospheric
and OceanicSciences,Universityof Michigan, Ann Arbor,
M148109.
15øS on the East Pacific Rise. Science, 214, 13-18,
1981.
Lyle, M. Estimationof hydrothermalmanganese
inputto the
oceans. Geology.4. 733-736, 1976.
(Received27 February1987;
accepted13 April 1987.)