intraplate volcanism: seamounts and...
TRANSCRIPT
13. プレート内火成活動：海山と海台 Intraplate volcanism: seamounts and plateaus
海洋底ダイナミクス 2018Ocean Floor Geodynamics 2018
プレート内火成活動 Intraplate volcanism Type1 古典的なホットスポット：不動か？“Classic” type hot spot Type2 スーパースウェル Super swell
Type3 リソスフェアのクラック＋プチスポット Lithosphere cracks + petit spot
巨大火成岩岩石区Large Igneous Provinces (LIPs) 巨大海台 Oceanic plateaus 環境・地球史への影響 Environmental impact
1
地球の火成活動（現世） volcanism on earth
(Press, Understanding Earth, 2003�
2
プレート内火成活動についての最近の認識
816 NATURE GEOSCIENCE | VOL 4 | DECEMBER 2011 | www.nature.com/naturegeoscience
FEATURE | FOCUS
Mantle plumes persevereAnthony A. P. Koppers
The ocean floor is littered with hundreds of thousands of mostly extinct volcanoes. The origin of at least some of these seamounts seems to rest with mantle plumes.
A recent census suggests that seamounts1 — typically extinct underwater volcanoes — are
numerous. It has been estimated that about 125,000 seamounts with a height of more than one kilometre exist on our ocean floors. Most of these are postulated to form at volcanic hotspots that are the surface expressions of mantle plumes — hot material upwelling from Earth’s interior. Yet, many seamounts do not show the typical characteristics expected for volcanoes that have formed above a mantle plume. So, debate about the feasibility of the mantle plume hypothesis is ongoing. The most straightforward explanation is that not all hotspot volcanoes are alike, and that some groups of seamounts are better explained by mechanisms other than mantle plumes.
Morgan’s mantle plumesSome hotspots form seamount trails along the surface of tectonic plates, far away from their volcanically active plate boundaries. Forty years ago, W. Jason Morgan introduced the concept of mantle plumes to explain this kind of hotspot volcanism2,3. According to his theory, plumes of hot material upwell from the deep mantle. During ascent and on impact with the overlying tectonic plates, these plumes drive melting and the production of magma, which erupts to form volcanic seamounts at the plate surface. Morgan proposed that the migration of tectonic plates over stationary and long-lived mantle plumes would generate chains of volcanoes on the ocean floor. However, because mantle plumes themselves cannot be directly sampled and the thin plume
conduits are difficult to resolve using seismic data4,5, their existence has been difficult to confirm. Many question whether all hotspot volcanism is formed by mantle plumes6,7 and some doubt whether mantle plumes exist at all8,9.
Perhaps the most captivating aspect of Morgan’s plume model is that he could explain the formation of the Hawaiian–Emperor and three other seamount trails along the Pacific Ocean floor. These three trails track each other in such a way that they can be explained by the rotation of a rigid Pacific plate that is drifting over four plumes fixed in the mantle (Fig. 1a). With this observation, Morgan supplied compelling support for the existence of plumes and also provided independent proof for the motion of tectonic plates relative to the underlying mantle. The mantle plume model also opened up potential new avenues of research into the Earth’s deepest regions10. Specifically, if the volcanic seamounts are the surface expression of a mantle plume, their erupted lavas could potentially preserve a record of long-lived variations in mantle composition and could provide insights into mantle convection.
Holes in the theoryThe plume model calls on an extensive global network of long-lived stationary mantle plumes that are continually delivering hot material from deep in the Earth. However, such a global network never fully materialized. It now seems that there aren’t many active hotspot systems around the world, maybe a few dozen — too few to have produced all of the world’s seamounts.
Of those seamounts that do seem to have formed above a mantle plume, some show evidence that the underlying plume was neither long-lived nor stationary. Improved mapping of seamount trails using satellite altimetry reveals that most seamount trails have typical life spans of just 30 million years11. And samples of lava collected from the Emperor seamounts during the Deep
Mid-oceanspreadingcentre
CMB
670 km
Hotspottype 3
Hotspottype 2
Hotspottype 1
Figure 1 | Models of ocean-island and seamount-trail formation: Courtillot’s framework6 of three hotspot types. The first, a classical Morgan-style long-lived mantle plume, originates from as deep in the mantle as the core–mantle boundary (CMB). The second hotspot type includes short-lived, smaller plumes originating from shallower parts of the mantle, probably as offshoots from large superplumes. These secondary hotspots are more common. The third type of hotspot is not related to any kind of mantle plume and may form where the oceanic lithosphere cracks or extends. This kind is the least investigated and may overlap considerably with the other hotspot types.
© 2011 Macmillan Publishers Limited. All rights reserved
(Koppers, 2011�
intraplate volcanism
3
ウィルソンの定義したホットスポット definition of hotspot by Wilson (1963)
relatively small, long-lasting, and hot regions -- called hotspots -- must exist below the plates that provide localized sources of high heat energy (thermal plumes) to sustain volcanism.
(Wilson, 1963�
4
固定ホットスポット仮説と海山列 fixed hotspot hypothesis and hotspot track
(Understanding Earth, 2003)
• Hotspots are fixed in deep mantle = no relative motion among hotspots
• Seamounts chain extending from hotspot is a kind of “track” of plate motion.
• support ‘plate tectonics’
• provide past plate motions
5
ハワイー天皇海山列 Hawaii hotspot and Emperor seamount chain
(Press, Understanding Earth, 2003�
6
ホットスポットの分布 Global distribution of hotspot
Based on UTIG hotspot list
7
ホットスポット火成活動の特徴 Hotspot volcanism : Features
• Courtillot et al.(2003)’s five criteria
• long-lived tracks
• traps at their initiation
• magma flux > 103 kg/s
• High 3He/4He or 21Ne/22Ne
• anomalously low shear velocities (Vs) in the mantle below
8
ホットスポット軌跡 track and mantle plume
(Press, Understanding Earth, 2003�
Oceanography Vol.23, No.144
(more or less linear) seamount trails that form as the plates constantly move over the “fixed” loci of the upwelling mantle plume stems (Richards et al., 1989).
Many observations are consistent with the existence of mantle plumes. Large mid-plate topographic swells (typically 1500–3000-km wide, up to
1500-m high, and correlating with long-wavelength gravity and geoid anomalies) have been found at the leading edges of many active seamount trails. The correla-tion implies that the swells are supported at depth by low-density subcrustal mantle material. This large-scale warping of otherwise rigid lithosphere is most
noticeable for the Hawaiian-Emperor seamount trail (e.g., Watts, 1976) and is believed to be directly related to the buoyancy of a plume and its interac-tion with the overlying Pacific Plate (Figure 1). Mapping of these swells using satellite-derived gravity and geoid data, for instance, allowed scientists to equate the sizes of these swells to vertical plume fluxes. Although the volume of active intraplate volcanism is small compared to island arc volcanism and the formation of the oceanic crust at the mid-ocean ridges, plume fluxes ranging from 1.0 Mg s-1 (Canary) to 8.7 Mg s-1 (Hawai`i) become significant when integrated over geological time and including all known hotspot systems (Davies, 1988; Sleep, 1990). Further observations that support the presence of mantle plumes include evidence that these lithospheric swells diminish away from active hotspots, the formation of linear age-progressive seamount trails, and the volcanic extinction of seamounts when plate motions move them away from their hotspot locations.
But mantle-plume behavior is not quite so simple, as the latest numerical mantle convection models suggest that a simple density-driven upwelling (Figure 1) is very unusual (but not implausible) and that the resulting plumes mostly are not vertically straight, narrow, and continuous, but often
Anthony A.P. Koppers ([email protected]) is Associate Professor, College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR, USA. Anthony B. Watts is Professor of Marine Geology and Geophysics, Department of Earth Sciences, University of Oxford, Oxford, UK.
23.6 Myr15.2 Myr7.0 Myr4.2 Myr1.4 Myr0 Myr
0 1Thermal Anomaly
FLEXURALMOAT
23 6 Myr15 2 Myr
MID-PLATESWELL
Figure 1. The Hawaiian-Emperor hotspot trail is our textbook example of the classical mantle plume model explaining the formation of intraplate seamounts. In this map of the Northwest Pacific (D. Sandwell and W.M. Smith: Gravity Anomaly Map based on Satellite Altimetry, Version 15.2), this archetypical seamount trail is exemplified by a deep flexural moat along its entire length and a significant mid-plate swell only prevalent toward the young southeastern end. The linearity of this seamount trail, in combination with a large mid-plate swell and a systematic age progression (with radiometric ages increasing toward the older northwestern end; see Figure 2), provides strong evidence for the existence of a mantle plume, maybe origi-nating deep in the mantle from a thermal anomaly (see simulation at the bottom by Van Keken [1997]). In this model, the seamount trail only forms after the plume head has dissipated and the narrow plume stem starts interacting with the lithosphere. Because the older Emperor seamounts have all been subducted into the Aleutian trench to the north, the fate of the plume head and any link to large igneous province volcanism are unidentified.
(Van Keken, 1997�
9
洪水玄武岩 Flood basalt
玉木、岩波地球惑星科学
10
地球化学的な特徴 Geochemical (isotopic) features
Anderson et al. http://www.mantleplumes.org/HeliumFundamentals.html
primordial isotope alpha decay of U and Th accumulate over time
typical values for R/Ra MORB 8+-2 OIB 5~42 Continent << 1
11
マントルプルームの深部構造 Deep structure of mantle plume
(Wolf et al., 2009)
•Dense network of Ocean Bottom Seismometers •low-velocity zone extends to lower mantle = mantle plume from D”?
12
Tab
le1
Scores
for49
hotspo
tswith
respectto
¢vecrite
riaused
todiagno
seapo
tentially
deep
origin
(see
text)
Hotspot
Lat
Lon
Track
Flood
/plateau
Age
Buo
y.Reliab.
3 He/
4 He
Tom
oCou
nt(‡E)
(Ma)
(500)
Afar
10N
43no
Ethiopia
301
good
high
slow
4Ascension
8S346
nono
/na
nana
00+
?Australia
E38S
143
yes
no/
0.9
fair
na0
1+?
Azores
39N
332
no?
no/
1.1
fair
high
?0
1+?
Baja/Gua
dalupe
27N
247
yes?
no/
0.3
poor
low
00+
?Balleny
67S
163
nono
/na
nana
00+
?Bermud
a33N
293
nono
?/
1.1
good
na0
0+?
Bou
vet
54S
2no
no/
0.4
fair
high
01+
?Bow
ie53N
225
yes
no/
0.3
poor
naslow
2+?
Cam
eroo
n4N
9yes?
no/
nana
na0
0+?
Canary
28N
340
nono
/1
fair
low
slow
2CapeVerde
14N
340
nono
/1.6
poor
high
02
Caroline
5N164
yes
no/
2poor
high
03
Com
ores
12S
43no
no/
nana
na0
0+?
Crozet/Pr.Edw
ard
45S
50yes?
Karoo
?183
0.5
good
na0
0+?
Darfur
13N
24yes?
no/
napo
orna
00+
?Discovery
42S
0no
?no
/0.5
poor
high
01+
?Easter
27S
250
yes
mid-Pac
mnt?
100?
3fair
high
slow
4+?
Eife
l50N
7yes?
no/
nana
na0
0+?
Ferna
ndo
4S328
yes?
CAMP?
201?
0.5
poor
na0
0+?
Galapagos
0268
yes?
Carribean
?90
1fair
high
02+
?Great
Meteor/New
Eng
land
28N
328
yes?
no?
/0.5
poor
na0
0+?
Haw
aii
20N
204
yes
subducted?
s80
?8.7
good
high
slow
4+?
Hog
gar
23N
6no
No
/0.9
poor
naslow
1Iceland
65N
340
yes?
Greenland
611.4
good
high
slow
4+?
JanMayen
71N
352
no?
yes?
/na
poor
naslow
1+?
Juan
deFu
ca/Cobb
46N
230
yes
no/
0.3
fair
naslow
2+?
Juan
Fernandez
34S
277
yes?
no/
1.6
poor
high
02+
?Kerguelen(H
eard)
49S
69yes
Rajmahal?
118
0.5
poor
high
02+
?Louisville
51S
219
yes
Ontong-Java
122
0.9
poor
naslow
3+?
LordHow
e(Tasman
East)
33S
159
yes?
no/
0.9
poor
naslow
1+?
Macdonald
(Cook-Austral)
30S
220
yes?
yes?
/3.3
fair
high
?slow
2+?
Marion
47S
38yes
Mad
agascar?
88na
nana
01+
?Marqueses
10S
222
yes
Shatski?
???
3.3
nalow
02+
?Martin
/Trind
ade
20S
331
yes?
no/
0.5
poor
nafast
0+?
Meteor
52S
1yes?
no/
0.5
poor
na0
0+?
Pitc
airn
26S
230
yes
no/
3.3
fair
high
?0
2+?
Raton
37N
256
yes?
no/
nana
naslow
1+?
Reunion
21S
56yes
Deccan
651.9
poor
high
04
StHelena
17S
340
yes
no/
0.5
poor
low
01
Samoa
14S
190
yes
no?
14?
1.6
poor
high
slow
4Sa
nFelix
26S
280
yes?
no/
1.6
poor
na0
1+?
Socorro
19N
249
nono
/na
poor
naslow
1+?
Tahiti/Society
18S
210
yes
no/
3.3
fair
high
?0
2+?
Tasmanid
(Tasman
central)
39S
156
yes
no/
0.9
poor
naslow
2Tibesti
21N
17yes?
no/
napo
orna
00+
?Tristan
37S
348
yes
Parana
133
1.7
poor
low
03
Vem
a33S
4yes?
yes?
(Orang
eR.)/
napo
orna
00+
?Yellowstone
44N
249
yes?
Colum
bia?
161.5
fair
high
02+
?
EPS
L6470
3-1-03
Cyaan
Magenta
Geel
Zwart
V.Courtillot
etal./Earth
andPlanetary
Science
Letters
205(2003)
295^308
297
13
Courtillot’s primary seven
Iceland
Afar
Reunion
Tristan
Louisville
Easter
Hawaii
Type 1
14
ホットスポットは移動するのか？ Does hotspot move?
Hawaiian hotspot track is predicted from a global plate motion chain based on relative plate motion data, and it is assumed that the Hawaiian hotspot is fixe relative to African hotspots, it does not fit the observed track
Obviously, it is also possible to construct models with substantiallylarger hotspot motion (for example, with lower viscosity, and/orlarger density anomalies in the mantle15); however, such models arenot consistent with observations, such as hotspot tracks, whereasthe models just summarized are broadly consistent. The directionsof hotspotmotion are less sensitive tomodel differences than are thespeeds of motion.
Relative plate motions and hotspot tracksWe now account for predicted hotspot motion and compute hot-spot tracks from specified relative plate motions. We cannot giveformal uncertainties, although—on the basis of the spread of modelresults as just discussed—we can judge whether a discrepancy seemsto be significant. The global tectonics of the Earth is characterized bya group of plates that diverge from the Indian and Atlantic oceans,and a Pacific group of oceanic plates that converge around most of
the Pacific rim. The relative motions of the two groups can bedetermined from observations of the sea floor in the South Pacificand southern Indian Ocean, with the caveat that deformation isknown to have occurred in continental regions of Antarctica andNew Zealand, but this is hard to quantify.For times younger than chron 20 (43Myr ago), substantial
Australia–Pacific plate motion (through New Zealand) cannot bedetermined with sufficient precision from local data for it to beuseful in this analysis. However, seafloor spreading in the SouthPacific is accurately quantified18, andmotion between East andWestAntarctica has been determined from a consideration of seafloorgeometry19. The motion of Africa, relative to East Antarctica sincethe Late Cretaceous, is determined from sea floor in the SouthwestIndian Ocean20. Hence the motion of Africa relative to the Pacificplate is determined by a plate motion chain running from AfricathroughEastAntarctica andWestAntarctica to thePacific (Figs 3, 4).
Figure 2 Computed hotspot motion and tracks for the moving-source model. Hotspotmotion is shown as rainbow-coloured lines (right colour bar), tracks as single-coloured
lines (shown for the past 83Myr); tickmark interval is 10Myr for both. Hotspot tracks are
computed for the motion of the plate where they are located, except for Hawaii, where
tracks computed for motion of the Pacific plate are plotted beyond its boundary. Tracks
are plotted regardless of whether a hotspot was actually beneath a given plate at a given
time. Red lines (model 1, shown for Hawaii and Louisville), and the purple line (model 2,
shown for Hawaii), are computed for absolute plate motions such that the fit to Tristan and
Reunion hotspot tracks is optimized for the two plate motion chain models as indicated;
optimization parameters are African plate rotations 0–47 and 47–83Myr ago. Green lines
(shown for Reunion and Tristan) are for optimizing the fit to Hawaii and Louisville for plate
motion chain model 1; parameters are Pacific plate rotations 0–25, 25–47, 47–62 and
62–83Myr ago. Black (model 1) and blue (model 2) lines are for optimizing jointly to all
four tracks for the two plate motion chain models as indicated; parameters are African
rotations 0–47, 47–62 and 62–83Myr ago. Dotted lines, hotspots assumed fixed; short-
dashed line, only hotspot motion on African hemisphere considered (shown for Hawaii);
long-dashed line, only hotspot motion on Pacific hemisphere considered (shown for
Hawaii); continuous lines, all hotspot motions considered. A least-squares method7 is
used to optimize the fit to locations and radiometric ages (see Fig. 5) of seamounts. Free
air gravity44 shows actual hotspot tracks. This, rather than topography, was chosen
because it should better distinguish between actual hotspot tracks (that is, seamounts
formed above a plume, in an intraplate location) and material erupted at a ridge through
plume–ridge interaction: the former are not expected to be in isostatic equilibrium and
should therefore be readily visible on a gravity map. The latter is expected to be
approximately in isostatic equilibrium, hence without strong gravity signature.
articles
NATURE |VOL 430 | 8 JULY 2004 | www.nature.com/nature 169© 2004 Nature Publishing Group
Steinberger et al.(2004)
15
ハワイホットスポットの南進 southward motion of Hawaiian plume
have been less frequent at Site 884 relative tothe other sites because of its flank position.The Site 1204 (Hole B) lavas record a lowangular dispersion, but these rocks might car-ry a CRM, explaining the agreement of theirmean inclination with that of the Site 884basalt section and the Site 1203 sediments.Because of potential inclination shallowing,the mean inclination from the Site 1203 sed-iments should be a minimum. This suggestsfurther that the mean derived from the basaltsat the same site is shallower because theavailable lavas underrepresent higher inclina-tion values.
We consider two scenarios: one in whichthe paleomagnetic results from the Site 884basalts, 1204 (Hole B) basalts, and Site 1203sediments best represent the field (paleolati-tude model A; IT ! 56.5° "12.6°
#12.4°, N ! 3) andanother in which we combine all the individ-ual basalt inclination units from Detroit Sea-mount into a mean (Model B; IT ! 52.9°"6.9°#3.7°, N ! 32). With either model, the paleo-latitude and age data yield average rates(Model A: 57.7 $ 19.2 mm year"1; Model B:43.1 $ 22.6 mm year"1) that are consistentwith the hypothesis that the Hawaiian hotspotmoved rapidly southward from 81 to 47 Ma(10). The values are consistent with updatedestimates of hotspot motion based on inde-pendent relative plate motions (5). Both pa-leolatitude models suggest that most of themotion occurred before the time of theHawaiian-Emperor bend (Model A, %44 Ma;Model B, %43 Ma). This is further supportedby the mean paleolatitude value from KokoSeamount (based on both thermoremanentmagnetizations and CRMs), which is only2.5° north of the fixed-hotspot prediction.Modeling of hotspot motion. Crust
ages available from marine magnetic anom-alies, radiometric age data from drill sites,and geochemical data (22) indicate that theHawaiian hotspot was close to a spreadingridge during the formation of Detroit Sea-mount (23). Hence, asthenospheric channel-ing of the plume (24) from a position to thesouth toward a more northerly ridge couldhave played some role in the differencebetween the paleomagnetic data and theprediction of the fixed-hotspot model. Themonotonic age progression of lavas recov-ered from Detroit to Koko Seamounts,however, leads us to believe that this po-tential channeling of plume material waslimited to the region at or north of DetroitSeamount. Furthermore, the similarity ofthe Hawaiian-Emperor chain with the Lou-isville chain of the South Pacific suggeststhat asthenospheric channeling was not thesole cause of the paleolatitude progression.
We examined whether the observed pa-leolatitude motion can be explained by ageodynamic model of the interaction of aplume with large-scale mantle flow. The flow
calculation (25) requires a mantle density andviscosity model and a surface-velocityboundary condition (13). A tomographymodel was used to infer mantle density vari-ations (26), and a viscosity structure based onan optimized fit to the geoid (with additionalconstraints from heat flow) (27) was applied.Both moving- and fixed-plume sources (28,29) that originate at the top of the low-viscosity layer at the base of the mantle wereconsidered (13).
Fast motion occurs when a conduit issheared and tilted in the large-scale flow andthe tilted conduit rises to the surface aided by
a large-scale upwelling. Thus, in these mod-els, fast hotspot motion corresponds to slowermantle flow rates (&10 to 20 mm year"1).Most computations (25) yield a hotspot mo-tion of 5° to 10° toward the south to southeastduring the past 100 million years (My). It ispossible to achieve a good fit to the paleo-magnetic data, because the age of the initia-tion of the Hawaii hotspot is unknown (andcan hence be used as a free parameter). Forthe moving-source model, southward motiontends to be faster if an earlier plume origin isassumed. However, plume initiation agesfrom 180 to 120 Ma (which imply the sub-
Fig. 3. (A) Paleolatitude data from ODP Leg 197 sites (1206, Koko Seamount; 1205, NintokuSeamount; 1204B and 1203, Detroit Seamount), ODP Site 884 (Detroit Seamount) (10), and DSDPSite 433 (Suiko Seamount) (11). Orange, results of thermal demagnetization; blue, results ofalternating field demagnetization. Result from 433 is based on AF and thermal data. Magenta,magnetization carried by hematite from weathered basalt from Site 1206. (B) Average paleolati-tude value for Detroit Seamount (square), based on inclination groups derived from basalts of Sites884, 1203, and 1204B (Model B, see text), plotted with select values from other seamounts (A).Also shown is a least-squares fit to the data (orange) and several paleolatitude trajectoriesrepresenting combinations of plate and hotspot motion.
R E S E A R C H A R T I C L E S
www.sciencemag.org SCIENCE VOL 301 22 AUGUST 2003 1067
The Emperor Seamounts: SouthwardMotion of the Hawaiian HotspotPlume in Earth’s Mantle
John A. Tarduno,1* Robert A. Duncan,2 David W. Scholl,3
Rory D. Cottrell,1 Bernhard Steinberger,4
Thorvaldur Thordarson,5 Bryan C. Kerr,3 Clive R. Neal,6
Fred A. Frey,7 Masayuki Torii,8 Claire Carvallo9
The Hawaiian-Emperor hotspot track has a prominent bend, which hasserved as the basis for the theory that the Hawaiian hotspot, fixed in thedeep mantle, traced a change in plate motion. However, paleomagnetic andradiometric age data from samples recovered by ocean drilling define anage-progressive paleolatitude history, indicating that the Emperor Sea-mount trend was principally formed by the rapid motion (over 40 milli-meters per year) of the Hawaiian hotspot plume during Late Cretaceous toearly-Tertiary times (81 to 47 million years ago). Evidence for motion of theHawaiian plume affects models of mantle convection and plate tectonics,changing our understanding of terrestrial dynamics.
The concept of an age-progressive set ofvolcanic islands, atolls, and seamounts pro-duced by a hotspot plume fixed in the deepmantle was first developed to explain theHawaiian Islands (1). The bend separatingthe westward-trending Hawaiian islandchain from the northward-trending Emper-or Seamounts has most often been inter-preted as an example of a change in platemotion recorded in a fixed-hotspot frame ofreference (2).
However, global plate circuits suggestlarge relative motions between Hawaii andhotspots in the Atlantic and Indian Oceans(3– 6 ). Improved mapping of marine mag-netic anomalies in the Pacific has failed to
define the directional change at 43 millionyears ago (Ma) (7 ) that would be expectedif such a large change in plate motion hadoccurred. There was also a general lack ofcircum-Pacific tectonic events (8) docu-mented for this time. Recent age data sug-gest a slightly older age for the bend [!47Ma (9)], but this revised timing still doesnot correspond to an episode of profoundplate motion change recorded within thePacific basin or on its margins.
One approach to examine hotspot fixityis to determine the age and paleolatitude ofvolcanoes that form a given hotspot track.For the Hawaiian hotspot, the paleolati-tudes of extinct volcanic edifices of the
Emperor chain should match the present-day latitude of Hawaii (!19°N) if the hot-spot has remained fixed with respect toEarth’s spin axis. The most reliable indica-tors of paleolatitude are basaltic rocks, buttheir reliability depends on each sectionspanning enough time to sample geomag-netic secular variation. Recovery of suchsamples requires ocean-drilling technology,and only a few seamounts have been sam-pled to date.
Paleomagnetic analyses of 81-million-year-old basalt recovered from Detroit Sea-mount (Site 884) yielded a paleolatitude of!36°N (10), which is discordant withHawaii. Data from !61-million-year-oldbasalt (9) from Suiko Seamount define apaleolatitude of 27°N (11). These data sug-gest that the Emperor Seamounts recordsouthward motion of the hotspot plume inthe mantle (10).A paleomagnetic test. The Ocean Dril-
ling Program (ODP) Leg 197 (12) sought to testthe hypothesis of southward motion of the Ha-waiian hotspot by drilling additional basementsites in the Emperor chain (Fig. 1). We collect-ed detailed stepwise alternating field (AF) de-magnetization data aboard the drilling shipJOIDES Resolution. Although these shipboarddata are of high resolution, they alone are in-sufficient to define paleolatitudes. Magneticminerals with intermediate to high coercivities,carrying magnetizations resistant to AF demag-netization, are commonly formed during sub-aerial or seafloor weathering. The magnetiza-tions of these mineral phases are easily resolv-able in thermal demagnetization data, which wealso discuss here (13).
The geomagnetic field at a radius r, co-latitude ", and longitude # can be describedby the gradient of the scalar potential ($):
1Department of Earth and Environmental Sciences,University of Rochester, Rochester, NY 14627, USA.2College of Oceanic and Atmosphere Science, OregonState University, Corvallis, OR 97331–5503, USA.3Geophysics Department, Stanford University, Stan-ford, CA 94305, USA. 4Institute for Frontier Researchon Earth Evolution, Japan Marine Science and Tech-nology Center, Yokosuka 237–0061, Japan. 5Depart-ment of Geology and Geophysics–School of Oceanand Earth Science and Technology, University of Ha-waii, Honolulu, HI 96822, USA. 6Department of CivilEngineering and Geological Sciences, University ofNotre Dame, Notre Dame, IN 46556, USA. 7Depart-ment of Earth, Atmospheric and Planetary Sciences,Massachusetts Institute of Technology, Cambridge,MA 02139, USA. 8Department of Biosphere-Geosphere System Science, Okayama University ofScience, Okayama 700–0005, Japan. 9Department ofPhysics, Geophysics Division, University of Toronto,Mississauga, ON L5L1C6 Canada.
*To whom correspondence should be addressed. E-mail: [email protected]
Fig. 1. Hawaiian-Emperor chain shownwith ODP Leg 197 sites (12) and marinemagnetic-anomaly identifications (40).
RESEARCH ARTICLES
22 AUGUST 2003 VOL 301 SCIENCE www.sciencemag.org1064
duction of volcanic edifices older than theoldest extant seamount, Meiji Guyot) gener-ally yield the best fits. For the fixed-sourcemodel, the computed hotspot motion consistsof two distinct phases. During the first phase,which lasts 100 to 150 My, southward motioncan be rapid. The second phase begins whenthe first conduit elements that arise from thefixed source reach the surface. The computedhotspot motion is slow during the secondphase and, in the example shown (Fig. 4),somewhat toward the north (13).
Overall, the results of the large-scale flowmodeling approach described above are con-sistent with the Leg 197 paleomagnetic data.Potentially important differences lie in thetotal motion predicted since !80 Ma (13) andin the need to incorporate in the modelingresults a change in plate motion at or near thetime of the bend. The paleomagnetic data donot require a change in plate motion, al-though a small change is not excluded.Implications of hotspot motion. The
hotspot motion defined by the new paleomag-netic and radiometric age data has implica-tions for a wide variety of issues, includingtrue polar wander (TPW) (30), the morphol-ogy of the past geomagnetic field, and thehistory of plate motions. Some investigators(31) have proposed that as much as 30° ofTPW (rotation of the entire solid Earth) hasaccumulated during the past 200 My. How-ever, a fixed-hotspot reference frame is usedto define TPW in these studies. The datapresented here, together with other tests (32,33), indicate that TPW has been overestimat-ed; Earth has been relatively stable with re-
spect to the spin axis since the Early Creta-ceous Epoch. Similarly, some changes in themorphology of the geomagnetic field withtime (34) that have relied on fixed hotspotsto anchor data from global sites are proba-bly artificial. One recent analysis that hasnot relied on the fixed-hotspot referenceframe has called for a significant axialoctopole contribution (g 0
3) to the time-averaged field (35). This conclusion is con-troversial, but if correct, it would implythat our paleolatitude calculations underes-timate the true hotspot motion.
Backtracking the position of early-Tertiaryand older Pacific basin sites, an essential aspectof some paleoclimate and tectonic studies, re-quires rethought, given that previous effortshave also relied on fixed hotspots. The north-erly position of the Late Cretaceous Hawaiianhotspot (23) casts doubt on the southern optionfor the Kula-Farallon ridge [a plate configura-tion that is typically called on to create highrates of northward transport for tectonostrati-graphic terranes in Alaska and British Colum-bia (36, 37)].
The fixed-hotspot interpretation of theHawaiian-Emperor bend implies that hugeplates can undergo large changes in directionrapidly. But such changes cannot be associ-ated with internal buoyancy forces (such assubducting slabs) because these require manymillions of years to develop. This has led tothe suggestion that plate-boundary forcesmight be responsible (38). The new paleolati-tude and radiometric age data (9) suggest thatchanges of plate motion at the time of theHawaiian-Emperor bend were much smaller
and more gradual than previously thought.Given the central role the Hawaiian-Emperorbend has played as an example of plate mo-tion change, these observations now raise thequestion of whether major plates can undergolarge changes in direction rapidly, and wheth-er plate boundary forces alone can play adominant role in controlling plate motion.
The similarity of the Hawaiian and Lou-isville hotspot tracks implies that the motionwe are tracking by the new paleomagneticdata is of large scale. This Late Cretaceous toearly-Tertiary episode of hotspot motion wasnot isolated; motion of the Atlantic hotspotsrelative to those in the Pacific occurred atsimilar rates during mid-Cretaceous times(39). These data sets indicate a much moreactive role of mantle convection in control-ling the distribution of volcanic islands. Attimes, it is this large-scale mantle convectionthat is the principal signal recorded by hot-spot tracks.
References and Notes1. J. T. Wilson, Can. J. Phys. 41, 863 (1963).2. W. J. Morgan, Nature 230, 42 (1971).3. P. Molnar, T. Atwater, Nature 246, 288 (1973).4. P. Molnar, J. Stock, Nature 327, 587 (1987).5. C. A. Raymond, J. M. Stock, S. C. Cande, in The Historyand Dynamics of Global Plate Motions, vol. 121,Geophysical Monograph Series, M. A. Richards, R. G.Gordon, R. D. van der Hilst, Eds. (American Geophysi-cal Union, Washington, DC, 2000), pp. 359–375.
6. V. DiVenere, D. V. Kent, Earth Planet. Sci. Lett. 170,105 (1999).
7. T. Atwater, in The Eastern Pacific Ocean and Hawaii,vol. N of The Geology of North America, E. L. Win-terer, D. M. Hussong, R. W. Decker, Eds. (GeologicalSociety of America, Boulder, CO, 1989), pp. 21–72.
8. I. O. Norton, Tectonics 14, 1080 (1995).9. W. D. Sharp, D. A. Clague, Eos 83, F1282 (2002).10. J. A. Tarduno, R. D. Cottrell, Earth Planet. Sci. Lett.153, 171 (1997).
11. M. Kono, Init. Rep. Deep Sea Drill. Proj. 55, 737(1980).
12. J. A. Tarduno et al., Proceedings of the Ocean DrillingProgram, Initial Report (Ocean Drilling Program, Col-lege Station, TX, 2002), vol. 197.
13. Material and methods are available as supportingmaterial on Science Online.
14. P. L. McFadden, A. B. Reid, Geophys. J. R. Astron. Soc.69, 307 (1982).
15. P. L. McFadden, R. T. Merrill, M. W. McElhinny, S. Lee,J. Geophys. Res. 96, 3923 (1991).
16. A. V. Cox, Geophys. J. R. Astron. Soc. 20, 253 (1970).17. J. A. Tarduno, Geophys. Res. Lett. 17, 101 (1990).18. W. A. Berggren, D. V. Kent, C. C. Swisher III, M.-P.Aubry, Soc. Econ. Paleontol. Mineral. Spec. Publ. 54(1995), pp. 129–212.
19. R. A. Fisher, Proc. R. Soc. London Ser. A 217, 295(1953).
20. G. B. Dalrymple, M. A. Lanphere, D. A. Clague, Init.Rep. Deep Sea Drill. Proj. 55, 659 (1980).
21. J. McKenzie, D. Bernoulli, S. O. Schlanger, Init. Rep.Deep Sea Drill. Proj. 55, 415 (1980).
22. R. A. Keller, M. R. Fisk, W. M. White, Nature 405, 673(2000).
23. R. D. Cottrell, J. A. Tarduno, Tectonophysics 362, 321(2003).
24. C. J. Ebinger, N. H. Sleep, Nature 395, 788 (1998).25. B. Steinberger, Geochem. Geophys. Geosyst. 3,10.1029/2002GC000334 (2002).
26. T. W. Becker, L. Boschi, Geochem. Geophys. Geosyst.3, 10.129/2001GC000168 (2002).
27. B. M. Steinberger, A. R. Calderwood, paper presentedat the European Union of Geosciences XI meeting,Strasbourg, France, 8 to 12 April 2001.
28. A. M. Jellinek, M. Manga, Nature 418, 760 (2002).
Fig. 4. (A) Computed Hawaiian hot-spot motion for fixed-source model(colored line), and tracks for fixed-source models (continuous line;plume initiation at 160 Ma) andmoving-source models (dashed line;plume initiation at 170 Ma). Tickmarkinterval is 10 Ma for both. (B) Com-puted changes of hotspot latitude forfixed-source plume model (13) (con-tinuous red lines) for plume-initiationages of 150, 160, and 170 Ma (upperto lower). Moving-source model (13)results (dashed purple lines) areshown for plume initiation at 180,170, and 160 Ma (upper to lower).Paleolatitude means for Koko, Nin-toku, Suiko, and Detroit Seamounts(Fig. 3) are also shown.
R E S E A R C H A R T I C L E S
22 AUGUST 2003 VOL 301 SCIENCE www.sciencemag.org1068
Tarduno et al.(2003)
Paleomagnetic and radiometric age data from samples recovered by ocean drilling define an age-progressive paleolatitude history.
16
天皇-ハワイ海山列の屈曲年代 Age of Emperor Hawaiian bend
Oceanography March 2010 47
knowledge about the motion of the plumes themselves as well as the best possible seamount geochronology.
Recent improvements in 40Ar/39Ar geochronology have allowed us to start addressing many of the above-described challenges in hotspot geodynamics and intraplate volcanism. Sensitivity improvements in mass spectrometry and the construction of low-blank extraction lines, in combination with aggressive application of incremental heating proto-cols, have allowed a renaissance in the age dating of seamounts (e.g., Koppers et al., 2003, 2007, 2008; Sharp and Clague, 2006; O’Connor et al., 2007). As a result, a more precise understanding of the timing of intraplate volcanism has
allowed us to accurately determine dura-tions of seamount formation (sometimes up to ~ 10 million years) and rates of age progression along seamount trails. Even though early geochronological studies squarely underwrote the hotspot model, later studies (for the same or other seamount trails, based on more extensive data sets, and using today’s analytical techniques) have revealed a significantly more complicated picture. For example, age dating and paleo-magnetic data from Ocean Drilling Program (ODP) Leg 197 demonstrated that the Hawaiian hotspot drifted south from ~ 35°N to 20°N between 80 and 49 million years ago (Figures 3 and 4; Tarduno et al., 2003; Duncan and
Keller, 2004; Duncan et al., 2006). The direction and magnitude of this drift is similar to recent modeling of plume advections within the context of whole mantle convection (Koppers et al., 2004; Steinberger et al., 2004) and suggests that initiation ~ 50 million years ago of the distinctive 120° Hawaiian-Emperor Bend (HEB) partially or even entirely reflects a change in the timing and magnitude of hotspot motion (Steinberger and O’Connell, 1998; Tarduno et al., 2003, 2009; Steinberger et al., 2004). Redating of the Louisville seamount trail (Koppers et al., 2004) has shown that a formerly “linear” age progression (Watts et al., 1988) is, in fact, nonlinear, with varia-tions in both hotspot and plate motions,
HAWAIIAN-EMPEROR BEND (HEB)
80
60
40
20
02000 1000 0
Pota
ssiu
m A
rgon
Age
(Ma)
Distance fromKilauea in Kilometers
NihoaFrench Frigate Shoals
Midway
Colahan
KokoOjin
Nintoku
Suiko
Detroit
Meiji
8.6 ± 0.2 cm/yr
Kauai
Tholeiitic LavaAlkalic Lava
Rejuvinated LavaPaleontologic Data
Topographic HighActive Volcanic VentLoci of Shield Volcanoes 160
20170
30
170
180
50
Kilauea LoihiSeamount
Niihau
French FrigateShoals
Necker
Nihoa KauaiOahu
MidwayColahanSeamount
HAWAIIAN-EMPEROR BEND (HEB)
HAWAIIAN ISLANDS
HAWAIIAN RIDGE
OjinSeamount
SuikoSeamount
EMPERO
R SEAM
OU
NTS
DetroitSeamount
NintokuSeamount
KokoSeamount
Miiji
Figure 3. Linear Hawaiian age progression derived from ages of Hawaiian-Emperor seamounts plotted against their distance from the active Kilauea Volcano based on data available from Clague and Dalrymple (1987), Duncan and Keller (2004), and Sharp and Clague (2006). Although the highly linear morphology of this seamount trail is more intricate when viewed close up (Jackson et al., 1972), these K/Ar and 40Ar/39Ar age data show a systematic (and more or less linear) aging of the shields of these volcanic islands and seamounts to the northwest and across the sharp 120° Hawaiian-Emperor Bend (HEB). (Koppers, 2010�
17
マントルに風が吹く？ Mantle wind model
particular model; however, comparison with our previouswork7,9,15–17 shows that these are typical results. By referring tothese previous results, we show that the conclusions here do notdepend on the choice of one particular set of model parameters butare robust conclusions that are valid as long as our model is at least
qualitatively correct. The computation consists of two steps. First, alarge-scale mantle flow field is computed, which is based on amantle density structure inferred from seismic tomography, globalplate motions and a radial mantle viscosity structure; this model fitsseveral observational constraints (see Supplementary Information1). The computed large-scale flow field is time dependent, althoughit does not change much over the period considered here. In thelower part of the mantle, the computed flow field is dominated bystructure of spherical harmonic degree two, with large upwellingsunder the Pacific Ocean and Africa, downwellings around thePacific Ocean, and flow towards the large upwellings in the lower-most mantle. In the upper part of the mantle, flow is also related toplate motions. In the uppermost part of the lower mantle, com-puted flow is a combination of outward flow away from the largeupwellings and plate return flow towards ridges. The latter might bean important contribution close to spreading ridges. Figure 1 showsa cross-section through density and flow field for different timesbeneath the Pacific Ocean.
In the second step, a plume conduit is inserted into the flow field.It is taken to be initially vertical. Plume initiation times are assumedto be the same as in ref. 7, except for Hawaii, where the age isunknown and 170 or 150Myr ago is used here. The base of theplume (assumed at depth 2,620 km, at the top of a low-viscositylayer) is assumed either to move with the horizontal component offlow or to be fixed in location (see Supplementary Information 1).The velocity of points along the plume conduit is computed as thevector sum of ambient mantle flow and a buoyant rising velocity(see Supplementary Information 1). It is assumed that plumeconduits do not influence larger-scale mantle flow. Hotspot surfacemotion is computed from the positions where, over time, the plumeconduit reaches the base of the lithosphere (assumed at depth100 km). Our choice of modelling parameters and assumptionshas been explained in previous work7,9,15,16.
Motion of plume conduits in a high-viscosity lower mantle isdominated by advection, and in a low-viscosity upper mantle it isdominated by buoyant rising. Consequently, hotspot motion tendsto be similar to the horizontal flow component at the depth at whichthe transition from low to high viscosity occurs (that is, the upperpart of the lower mantle)16. For the viscosity structure used, thehorizontal flow at this depth has a root-mean-square value of,1 cmyr21. Beneath the Tristan and Reunion hotspots it is domi-nated by outward flow from the large upwelling under Africa; in thecase of Louisville it is a combination of flow away from the largeupwelling under the Pacific Ocean and plate return flow towards thePacific–Antarctic ridge17. The computedmotion of these hotspots isindeed slow and is as expected from the flow pattern (Fig. 2). For theHawaiian plume, another effect leads to faster (a few cmyr21)hotspot motion. Figure 1 shows a projection of the Hawaiianconduit; it is strongly tilted in a north–south direction. Flow inthe upper part of the mantle is to the north and in the lower part tothe south, which tilts and distorts the plume. Subsequently, therising of a tilted plume conduit, aided by a large-scale upwelling,causes comparatively rapid hotspot motion. Among the hotspotsconsidered here, we find this effect only for the Hawaiian hotspot.
We have previously shown results for a larger number of models,and discussed the dependence of results on various model par-ameters7,9,15–17. In summary, these results yield the following. ForHawaii, the motion is in a southerly to southeasterly direction. Theaverage speed is moderately slow (,1 cmyr21), but episodes offaster motion (several cm yr21), lasting for several tens of Myr occurfor somemodels. Computed hotspot motion tends to be faster, if anolder age for the plume is assumed. For Louisville there is slowmotion (,1 cm yr21 or less) in various directions—in most cases ina southeasterly direction. For Reunion there is slow motion(,1 cmyr21 or less) in an easterly direction. For Tristan there iseither very slow motion (less than 1 cmyr21) in a southeasterly tosouthwesterly direction, or an almost stationary hotspot.
Figure 1 North–south mantle cross-section at 1558W. For different times, computedmantle density anomalies are shown in rainbow colours, and north–south and vertical
components of computedmantle flow as arrows. a, 120Myr ago;b, 90 Myr ago;c, 60Myrago;d, 30Myr ago;e, present. Also shown is the projection of the predicted Hawaiianplume conduit for source moving with the flow (plume initiation age 170Myr ago;thick red
line) and fixed source (age 150Myr ago;thick violet line) at those times.
articles
NATURE | VOL 430 | 8 JULY 2004 | www.nature.com/nature168 © 2004 Nature Publishing Group
Steinberger et al.(2004)
In the deep mantle beneath the northern Pacific, the mantle flows southward. Mantle plume from hotspot bend along the flow, causing southward rapid motion of Hawaiian hotspot and no rapid motion of Louisville hotspot.
ARTICLES NATURE GEOSCIENCE DOI: 10.1038/NGEO1638
Cou
ntC
ount
Cou
ntC
ount
Inclination
Site U1372n = 21 archiven = 9 discrete
Sites U1373and U1374
Site U1376
68°¬69°
2-cm archive-half
Discrete samples
0
50
100
150
Present-dayLouisville
hotspot
n = 28 archiven = 28 discrete
n = 8 archiven = 8 discrete
0
50
100
0
50
100
150
0
50
100
¬90° ¬80° ¬70° ¬60° ¬50° ¬40°
Site U1377n = 12 archiven = 3 discrete
Figure 4 | Bootstrap inclination distributions for individual Louisvilleseamounts drilled by IODP Expedition 330. Distributions for both 2-cmarchive-half (red) and discrete sample (blue) flow means are shown, eachrepresenting the bootstrap results of 1,000 resamplings with replacementand incorporating within-flow dispersions, using only units with ISCI = 3 or2. Inclination averages and 1� (filled rectangles) and 2� uncertainties(lines) exclude pseudosamples where inclination-only20 averaging failed toconverge (yielding inclinations of �90� that are excluded in theuncertainties of our average palaeolatitude estimates). The vertical greyline indicates the expected geocentric axial dipole inclination for thepresent-day hotspot location (50.9–52.4� S; refs 4,17,21,25).
Louisville seamounts is less straightforward than for sites inthe Hawaiian–Emperor seamount trail5,18,19. We developed anobjective, although qualitative, index (the in situ confidenceindex, ISCI; see Supplementary Information) to determine whetherindividual units were probably in situ, ranging from ISCI = 3 forunits that are definitely in situ to ISCI = 0 for those that areprobably not. Volcaniclastic units were assigned an ISCI value
of ‘not applicable’ because it is uncertain whether clasts in thesevolcanic sediments have retained their orientation since eruption.In this study we use results only from the most reliable lava flowunits and dikes (ISCI= 3 or 2).
Results from Rigil Guyot provide the best documented in-clination record, with 522m penetration at Site U1374 and anadditional 66m drilled at Site U1373, located ⇠10 km to the easton the summit plain. Site U1374 has dominantly steep negativeinclinations (normal polarity) with several reversed polarity flowsin the uppermost 45m and remarkably consistent inclinations inmany volcaniclastic breccias, similar to that of intercalated in situflows or dikes (Fig. 3). A total of 19 in situ flow unit means for SiteU1374 were determined from discrete samples (seven additionalunits were not sampled at sea). After correction for deviation ofthe borehole from vertical (2.2� correction; see SupplementaryInformation) these yield a mean inclination of �68.7± 8.4� us-ing inclination-only averaging20. An additional 9 flow units fromSite U1373 give a shallower mean inclination of �55.2± 10.6�,similar to moderate inclinations recorded at both the top andbase of U1374. As lava flows of Site U1373 erupted ⇠1.0Myrlater than lava flows in the deepest basement units of Site U1374,the shallower inclinations at Site U1373 are more likely to reflectpalaeosecular variation rather than Pacific Plate motion or drift ofthe Louisville hotspot.
We have therefore combined flow units from Sites U1373and U1374 to calculate an overall mean inclination for RigilGuyot at ⇠70Myr ago. For this and other guyots, we havetreated each flow unit as independent, because we were unable tosample all in situ flow units at sea and the remaining unsampledunits would probably affect identification of inclination groups18.More importantly, the presence of intercalated sediments andvolcaniclastics suggests that many flow units in fact representindependent samples of the geomagnetic field. Owing to the smallnumber of lava flows, we use a bootstrap resampling to provide themost robust estimate of the mean inclination and its uncertaintyfor each site (Fig. 4 and Table 1; see Supplementary Information fordetails). In contrast to earlier studies, we also explicitly incorporatewithin-flow inclination dispersion (with a median kappa of ⇠280for discrete sample flow means). The resulting distributions for2-cm archive-half measurements and discrete samples from RigilGuyot are similar and statistically indistinguishable (at the 1�confidence level) from the geocentric axial dipole inclination(±68�) for the present-day hotspot location at⇠51� S.
The resulting palaeolatitude for Rigil Guyot is 47.0� S± 8.0�
(n=28) and its distribution ofmean flow inclinations is statisticallysimilar to that expected from global geomagnetic field modelsat ⇠51� S (Fig. 5). This result alone suggests that the Louisvillehotspot experienced limited latitudinal motion since 70Myr ago.Therefore, assuming the simplest possible palaeolatitude history,it follows that the younger Louisville volcanoes are likely to havesampled the geomagnetic field at a similar latitude. Even thoughwe sampled only a limited number of flows from the 64- and50-Myr-old Burton and Hadar guyots, they each are statisticallyindistinguishable from 51� S at 49.8� S±4.8� (n= 9) and 52.3� S±20.2� (n = 3). It is unlikely that the results from either guyotadequately average palaeosecular variation, but nevertheless theseresults are broadly compatible with the palaeolatitude estimate ofRigil Guyot and, therefore, a limited Louisville hotspot motion, atleast since 70Myr ago.
Together, the inclinations from Sites U1373–U1377 representa relatively large number of flows, from one dominantly normalpolarity and two reversed polarity guyots, which may provide anadequate sampling of geomagnetic palaeosecular variation. To testthis, we compare the directional scatter (circular standard deviation,✓63) of the combined data set to two statistical models of thegeomagnetic field (CJ98, TK03; refs 22,23) that were designed
914 NATURE GEOSCIENCE | VOL 5 | DECEMBER 2012 | www.nature.com/naturegeoscience
Kopper et al.(2012)
Louisville IODP
18
スーパースウェル：フレンチポリネシア Super Swell: French Polynesia
chain at the scale of the alignment and this chainwill not be discussed any further.
4.2. Society
4.2.1. Volcanic Chain Description
[44] The Society islands (Figure 11a) are situatedbetween latitudes 16!S and 19!S and longitudes147!W and 153!W on a seafloor displaying agesbetween 65 and 95 Ma. They stretch along a200-km-wide and 500-km-long band orientated inthe direction of the present Pacific plate motion:N115 ± 15!. The age progression is uniformfrom the youngest submarine volcano, Mehetia(0.264 Ma [Duncan and McDougall, 1976]), situ-ated at the southeast extremity to the oldest datedisland, Maupiti (4.8Ma [White and Duncan, 1996]).
4.2.2. Swell
[45] The topographic anomaly associated with thisalignment is shown in Figure 11b. It stretchesalong the volcanic chain. Its maximal amplitude,980 m, is reached 30 km northwest of Tahiti, and is
not correlated with any volcanic structure. Thisdemonstrates that the volcanoes contribution isefficiently removed by the MiFil method. For thisvolcanic chain, the swell description correspondsto the one previously reported for hot spot swells,created by the simple interaction of a plume withthe lithosphere: the swell’s maximum is located215 kilometers downstream from the active volca-nism, the swell stretches along the volcanic chainand subsides along the direction of the platemotion.
4.2.3. Buoyancy Flux
[46] For the Society volcanic chain we find abuoyancy flux of 1.58 ± 0.15 Mg s!1, using VL =110 mm yr!1. Previous estimations [Davies, 1988;Sleep, 1990] are greater, mostly because theauthors overestimated the swell volume in notremoving completely the volcanoes.
[47] As pointed out by Courtillot et al. [2003],several criteria (plume duration, traps at theirinitiation, rare gas isotopic ratio, Vs at 500 kmdepth and the buoyancy flux) allow to establish the
Figure 10. Swell amplitudes (in meters). The general map represents the case in which two swells are found for theCook Islands. The case when one swell only is found is represented in the top left corner inset. The black line is the3000 m isobath.
GeochemistryGeophysicsGeosystems G3G3 adam et al.: seafloor swells 10.1029/2004GC000814
13 of 25
(Adam et al., 2005�
Type 2
水深が年齢に比して浅い water depth is shallower than expectedly age-depth curve
19
リソスフェアの裂け目から漏れる？ lithospheric cracking?
2003]. Gans et al. [2003] suggested instead that cracking ofthe lithosphere is caused by bending of the plate underthermal stresses associated with cooling of the lithosphere.Sandwell and Fialko [2004] showed that thermal stresseswould lead to a preferred wavelength of bending andcracking, which could reproduce the initial 150 to 200 kmwavelength of cross-grain gravity lineations [Haxby andWeissel, 1986]. Both of these variations on the tectoniccracking hypothesis are passive models in which the vol-canic ridges form in response to cracking; either the cracklets preexisting asthenospheric melt escape to the surface orthe small amount of local extension induces some upwellingthat causes pressure release melting.[4] Alternative models for the origin of the volcanic
ridges involve anomalous melting in the asthenosphereassociated with small-scale convection, minihot spots, orreturn flow of the asthenosphere to the East Pacific Rise thatincorporates compositional and/or thermal anomalies withinthe returning flow field. In these models, the volcanic ridgesare caused by active processes beneath the lithosphere thatcreate temperature variations and pressure-release melting.The ridge-like nature of the individual edifices could beattributed to dikes propagating under the influence of aremotely applied stress field or under the stresses associatedwith loading of the lithosphere by a linear chain of sea-mounts [Hieronymus and Bercovici, 2000].[5] One way to distinguish between lithospheric cracking
and asthenospheric models for the origin of the volcanic
lineations is to examine in detail the morphology of theseafloor and the distribution of eruptive centers and lavaflows. One could argue that the evidence for cracks preced-ing volcanism has been obscured by subsequent burialbeneath the built-up edifices, but if we look at the processin the early stages of formation, we should be able torecognize which type of model is more likely by searchingfor faults or linear features and by establishing whether theinitial volcanic activity is localized or widespread. The PukaPuka chain is now a fossil feature, except perhaps for somescattered volcanism in the Rano Rahi seamount field[Scheirer et al., 1996a, 1996b], but we have recentlymapped two other neighboring, linear volcanic features thatare still actively forming to the west of the East Pacific Rise.[6] This paper describes the distribution of recent intra-
plate volcanism and the morphology of the volcanic featuresin the Gravity Lineations Intraplate Melting Petrology andSeismic Expedition (GLIMPSE) study area, with an em-phasis on distinguishing between lithospheric cracking andasthenospheric origins for the volcanic activity. In this areajust to the north of the Rano Rahi seamount field (Figure 2),young lava flows and the Hotu and Matua seamounts werefirst discovered in the early 1990s during mapping inpreparation for the Mantle Electromagnetic and Tomogra-phy (MELT) Experiment. The existence of the SojournRidge was first noted on high-resolution images of satellitefree-air gravity anomalies [Smith and Sandwell, 1997], andit was partially mapped from shipboard for the first time in
Figure 1. Predicted bathymetry of the East Pacific Rise in the vicinity of the GLIMPSE study areabased on satellite altimetry [Smith and Sandwell, 1997]. The broad shallow area immediately north of theSojourn Ridge at about 116!W does not exist (see Figure 2); it is an artifact stemming from a geoid orgravity anomaly that does not have corresponding bathymetric expression. The inset map of the globeshows the location of the GLIMPSE study area in the box, with heavy black lines indicating plateboundaries.
B11407 FORSYTH ET AL.: GLIMPSE SEAMOUNTS AND RIDGES
2 of 19
B11407
Figure 2
B11407 FORSYTH ET AL.: GLIMPSE SEAMOUNTS AND RIDGES
3 of 19
B11407
Figure 3. (a) High-resolution bathymetry and (b) side-scan sonar images of Hotu Matua region(box Ain Figure 2). Matua is dominated by a single conical feature, while Hotu is a coalescence of at least fourflat-topped volcanic centers. The most reflective patches representing the roughest and presumablyyoungest seafloor are scattered around and to the east of Matua. An example of botroidal topography isfound to the east of the main edifice of Matua. Blue lines indicate older, less reflective, bathymetricfeatures that separate the Hotu and Matua recent flows and which probably formed nearer the East PacificRise (EPR); they are labeled as preexisting features. Dredge locations with Ar/Ar ages are shown as reddots with age determinations indicated. In this and subsequent detailed images, pixel size is 200 m.Artificial illumination of the bathymetry is from the east.
B11407 FORSYTH ET AL.: GLIMPSE SEAMOUNTS AND RIDGES
6 of 19
B11407
(Forsyth et al, 2006�
Type 3
20
プチスポット petit spot
Fig. 2
–5500–5500
–4000
–2000
–6000–8000
–6000–6000
–5500
Kuril Tren
ch
Kuril Tren
ch
Japa
n Tr
ench
Japa
n Tr
ench
Joban Joban SeamountsSeamounts
Japanese Japanese SeamountsSeamounts
Erimo Erimo SeamountSeamount
DaiichiDaiichi–Kashima Kashima SeamountSeamount
NosappuNosappuFracture Fracture ZoneZone
Pacific Plate
outer-rise
Japan Trench
Site A
Site B
Site ASite B
Fig.1. Bathymetricalmap of o¡shore NE Japanusing the data of Smith &Sandwell (1997)contoured at 500mintervals (upper panel)and a model for petit-spot related to the plate£exure (lower) modi¢edfromHirano et al. (2006).White and brown boxesshow the area of Fig. 2and the region of cross-sectional model shown bythe lower ¢gure,respectively.White heavylines show theCretaceous fracture zone(Nakanishi &Winterer,1998). Area of outer-riseis shown by hatched area,which is approximatelycorrespond to shallowerarea than 6000 metersbelow sea level (mbsl) for130^140Ma ocean £oor atthe western side offracture zone and than5900mbsl for120^125Maocean £oor at the easternside of the fracture zone,respectively.
r 2008 The AuthorsJournal Compilation r Blackwell Publishing Ltd, European Association of Geoscientists & Engineers and International Association of Sedimentologists544
N. Hirano et al.
age spectra for acid-leaching samples of altered submarinebasalts show high apparent ages and low Ca/K ratios forthe low temperature steps, and low apparent ages and highCa/K ratios for the high temperature increments. Accord-
ing toKoppers etal. (2000) this is due to the di¡erential de-gassing of alteration minerals at the low temperature stepsand the degassing of primary plagioclase and clinopyrox-ene phases at the high temperature steps, all in combina-tion with various redistributions between Ar isotopes dueto 39Ar and 37Ar recoil. These e¡ects may be more severewhen the glass content of the groundmass is high, butwell-developed age plateaus are generally formed at theintermediate temperature steps in case of holocrystallinegroundmass samples. We see similar systematics in ourgroundmass age spectra. However, two samples did notproduce clear age plateaus, and two other samples onlyformed narrow three-step age plateaus.This indicates thatalteration could not su⁄ciently be removed or that thegroundmass samples contained too much devitri¢ed ba-saltic glass.We thus need to be cautious in our interpreta-tion of these age spectra and only use our results todistinguish between samples of Cretaceous age and sam-ples that are younger than10Ma, the low age expected fortypical petit-spot volcanoes (Hirano et al., 2006).
The petit- spot basalt of 6K#880-R3B, on the otherhand, does show a well-developed age plateau that is 83%wide and includes eight heating steps. However, the2.54! 0.76Ma plateau age is too high due to excess 40Aras indicated by the 40Ar/36Ar ratio of 340 ! 21that is high-er than the atmospheric ratio (295.5) in the inverse iso-chron. The 1.76 ! 0.58Ma age from the inverse isochrondiagram thus appears the best age estimate for sample6K#880-R3B and is consistent with its morphologicaldesignation as a young petit- spot volcano.
DISCUSSIONTwo stages of intra-plate volcanism characterize the NWPaci¢c Plate spanning a geological time period from theEarly Cretaceous to the recent past. In the ¢rst stage, theseamounts of theWest Paci¢c Seamount Province (WPSP)and Joban and Japanese Seamount Trail (JJST) wereformed during the Late Jurassic and Cretaceous (Koppersetal., 2003).The new circular knolls described in this studyappear in the Late Cretaceous and thus seem to be part ofthis early stage of intra-plate volcanism. However, as wewill discuss below, their morphology may also indicatethat theywere formed as o¡-ridge volcanoes. In the secondstage, the irregularly shaped and low-volume petit- spotsappear, but only during the last 10Myr and only at loca-tions on the outer-rise of the subducting Paci¢c Plate (e.g.Hirano et al., 2006).This latter kind of intra-plate volcan-ism seems unrelated to the massive hotspot volcanism thatcharacterizes the West Paci¢c during the Cretaceous,instead it appears to be directly related to the subductionprocess and the bulging of the downgoing slab (e.g.Watts& Zhong, 2000). Here, we will further expand on the mor-phological characteristics and seamount ages that allow usto distinguish between the large seamounts formed in theCretaceous and the small petit- spot volcanoes formedmore recently in relation to the subducting Paci¢c Plate.
(a)
(b)
(c)
De pth (m)
(d)
Fig. 3. Bathymetrical maps (left ¢gures) and sidescan sonarimages (right ¢gures) of circular knolls shown each site in Fig. 2.Bathymetrical maps are contoured at100m intervals. Horizontalresolutions of bathymetrical and sidescan map are 300 and 50m,respectively.
r 2008 The AuthorsJournal Compilation r Blackwell Publishing Ltd, European Association of Geoscientists & Engineers and International Association of Sedimentologists 547
Seamounts, knolls and petit-spotmonogenetic volcanoes on the subductingPacific Plate
HIRANO ET AL.: A NEW TYPE OF INTRA-PLATE VOLCANISM 2721
Table 1. Bulk compositions of 10K#56 R-001 and R-002 by XRF analysis.
sample major element (wt%) SiO 2 TiO 2 AI203 Fe203 FeO MnO MgO CaO Na20 K20 P205 H20+ H20-
R-001 bulk 48.22 2.56 10.30 6.38 4.23 0.14 11.68 7.47 2.95 3.17 0.78 1.76 0.36
groundmass a 48.84 2.72 11.60 10.44 - 0.13 8.23 7.68 3.04 3.88 0.83 1.23 1.39 R-002 bulk 48.45 2.84 11.26 6.12 4.36 0.13 7.83 8.14 3.28 3.63 0.84 2.49 0.64
groundmass a 49.27 2.84 12.18 10.00 - 0.12 6.61 7.61 3.13 4.11 0.87 1.47 1.81 trace element (ppm)
Ba Ce Co Cr sample
Total
100.00
100.00 100.00 100.00
Ga Nb Ni Pb Rb Sr Th V Y Zr
R-001 bulk 1202.3 94.4 56.7 527.8 17•7 36.6 418.1 7.2 47.0 1076.5 2.0 140.1 groundmass d 1176.8 -154.0 424.0 - 37.4 163.1 6.9 56.2 1212.4 4.7 -
R-002 bulk 1209.5 108.7 46.7 400.5 20.2 40.3 227.8 7.3 53.1 1092.4 2.0 143.4
groundmass d 1191.2 - 129.0 353.0 - 39.4 117.2 7.3 59.0 1140.4 4.9 -
a Data for the samples separated the olivine phenocryst from the groundmass. In this data a, Fe203 show the total Fe-oxide.
14.2 243.0 18.8 263.8 15.3 259.9 19.7 273.8
Table 2. REE compositions of the bulk samples by the ICP-MS analysis. Analyzed by Dr. M. Komuro and Ms. K. Fujii, Institute of Geoscience, University of Tsukuba (personal communication).
ppm
Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu R-001 19.17 54.07 106.95 12.03 49.35 9.41 2.98 7.74 1.00 4.63 0.77 1.83 0.23 1.41 0.19 R-002 20.78 58.69 114.28 12.97 53.61 10.02 3.22 8.38 1.08 4.72 0.88 1.93 0.23 1.46 0.20
4. Tectonic interpretation and geophysical implication
The alkali-basalts documented here are very young, much younger than the ocean floor in this area (identified isochron M9 or M10; around 130 Ma [Gradstein et al., 1994; Kobayashi et al., 1998]). Seamounts in this region are also Cretaceous, with ages ranging from 120-104 Ma. In contrast, the rocks analyzed in this study have ages of 5.95+0.31 Ma (latest Miocene), and are found in small-volumes along seafloor escarpments rather than on seamounts or large volcanic constructions. By performing a plate tectonic reconstruction we have determined that there is no plausible hotspot that could have produced these basalts. We have reconstructed the eruption location of these basalts using the 40Ar-39Ar age of 5.95 + 0.31 Ma and the present "absolute" motion of the Pacific Plate (10.29 cm/yr to 295.26 degrees [ Gripp and Gordon, 1990]). Using this method we have derived a position of approximately 612+32 km ESE off the northern Japan Trench, now approximately at 37øN, 149øE. As the volcanic front in the NE Japan Arc has scarcely shifted since the latest Miocene [Ohki et al., 1993], the kinematics of the Pacific slab at 5-6 Ma are the same as at present. According to the
lOO
o.1 sr K RI:BaThTaNI:CeP ZrHfSmTi Y YbScCr LaCePr Nd SmEuGdT'bDyHoErTmYbLu
Figure $. Spidergrams of trace element concentrations in samples 10K#56 R-001 and R-002. A: Normalized by average MORB [Pearce, 1982, 1983]. B: Chondrite-normalized REE pattern [Evensen et al., 1978]. U, Ta and Hf were not analyzed.
bathymetric chart of the northwest Pacific, this area corresponds to a site just oceanward of the current outer swell or forebulge (Hokkaido rise), with an inferred paleo-depth of-6000 m (asterisk in Fig. 1A).
Enriched incompatible element concentrations and REE pattern indicate that the magma source for these alkali-basalts formed as a result of low degree of partial melting. Disequilibrium between olivine phenocrysts and groundmass olivines suggest that the phenocrysts may be xenocrysts transported from deep in the mantle with rapid rise of the alkali-basaltic magma. If a fracture occurs or is rejuvenated in the
03201.
q:10
Plateau age=5.95_0.31 Ma
39Ar=94 % (n=4)
A.
•3
0 o 39Ar cum. % 50 0.2 0.4 39Ar/4øAr 0.8 1.0
ß . ,
MSWD = 1 .O6 B 40 36 -I- 4 "',k (4 Ar/ Ar),=o= 304.9_9.
Age (Ma) lO 8 6
100
3
10•
Figure 6. Age determination results by 40Ar-39Ar method of sample 10K#56 R-002. Correction factors and J-value are as follows; (36Ar/37Ar)ca = (3.744-1-0.082) X 10 -4, (39Ar/37Ar)ca = (9.30+0.44)X 10-4 and J = (3.412+_.0.063)x 10-3. Calculated using decay constants and potassium isotope ratios from Steiger and Jager (1977).
GEOPHYSICAL RESEARCH LETTERS, VOL. 28, NO. 14, PAGES 2719-2722, JULY 15,2001
A new type of intra-plate volcanism; young alkali-basalts discovered from the subducting Pacific Plate, northern Japan Trench
N. Hirano l, K. Kawamura ?, M. Hattori 3, K. Saito 4 and Y. Ogawa -• Abstract. Alkali pillow basalts were collected from the toe of the oceanward slope of the northern Japan Trench. These alkali-basalts formed as a result of a low degree of partial melting of Pacific Ocean mantle and rapid rise of the magma (no fractionation in shallow magma chambers). Reconstructing Pacific Plate motion based on 40Ar-•9Ar age dates of 5.95_ 0.31 Ma for these basalts indicates that they erupted outboard of outer swell or forebulge of the Japan Trench in the NW Pacific. We suggest that these alkali-basalts represent a new form of intra-plate volcanism, whereby magmatic activity occurs off the forebulge of the downgoing Pacific slab, perhaps using conduits related to fracturing of the slab during bending prior to subduction.
1. Introduction
Alkali-basalts occur on various parts on the surface of the earth, most particularly in continental and hotspot areas. These alkali-basalts are products of deep-origin magma from the upper mantle or lower depths. Occurrences of such alkali-basalt are also documented from tectonically unique locations, such as along deep fractures in oceanic crust.
Alkali olivine basalt and trachyandesite representative of the ocean island basalt series have been documented around the Japan Trench on the Joban, Erimo and Takuyo Seamounts [Kobayashi et al., 1987; Cadet et al., 1987]. The 40Ar•9Ar ages of these volcanic rocks range from 120 Ma (Daiichi-Kashima Seamount in the Joban Seamount Chain) to 104 Ma (Erimo Seamount) [Takigami et al., 1989], indicating that these are the products of Cretaceous off-ridge seamount volcanism (Fig. 1A). In contrast, the bathymetry of the study area does not show any evidence for a large volcanic edifice or seamount, but only a small mound (Fig. lB).
This paper describes an occurrence of young alkali-basalt on the downgoing oceanic slab of a subduction zone. We present the geologic setting, major and trace element compositions, and 40Ar-39Ar age of these basaltic rocks. We then discuss the tectonic and geophysical implications for this first documentation of alkali-basaltic magmatism outboard of outer swell of a subducting oceanic slab.
2. Occurrence and description of samples Continuous outcrops of pillow basalt were documented and
sampled at depths of 7325 to 7360 m on the oceanward slope toe
IDoctral Program in Geoscience, University of Tsukuba, Tsukuba, Japan. Now at Ocean Research Institute, University of Tokyo, Tokyo, Japan.
2Fukada Geological Institute, Tokyo, Japan 3Japan Marine Science and Technology Center, Yokosuka, Japan 4Department of Earth and Environmental Sciences, Faculty of Science,
Yamagata University, Yamagata, Japan 5Institute of Geoscience, University of Tsukuba, Tsukuba, Japan
Copyright 2001 by the American Geophysical Union.
Paper number 2000GL012426. 0094-8276/01/2000GL012426505.00
of the northern Japan Trench (39023 ' N, 144016 ' E) during JAMSTEC (Japan Marine Science and Technology Center) R/V Kairei/ROV KAIKO cruise KR97-11. The slope is characterized by trench-parallel (N-S) normal faults with some NNW or NNE faults, due to warping of the downgoing Pacific Plate (the age of the Pacific Plate here is Early Cretaceous [Kobayashi et al., 1998]) (Fig. 1). These normal faults bound horst and graben structures that are approximately 5 km in horizontal extent with 100 to 500 m vertical separations [Ogawa and Kobayashi, 1993;
140øE 145'E 1501E
0%•
[ • •J.•_ Rise ,]
...........
Figure 1. Index maps of the dive site. A: The general bathymetric map of the northwest Pacific ocean floor based on Kobayshi et al. (1998). Black area is trench floor deeper than 7000 m, and grey-shaded area is the outer swell (<5400 m in depth). Radiometric ages of Erimo (a), Ryofu (b), Daini-Kashima (c) and Daiichi-Kashima (d) Seamounts are obtained as follows respectively; (a) 104 Ma and (d) 120 Ma 40Ar-39Ar age [Takigami et al., 1989], (b) 70-72 Ma and (c) 81 Ma K-Ar age [Ozima et al., 1977]. The approximate eruption site is plotted as asterisk. B: Seabeam bathymetric map of the dive site 10K#56 by R/V KAIREI at the northern Japan Trench. Contour interval is 250 m. Trench axis is shown by white arrow.
2719 (Hirano et al., 2001; 2006; 2008)
Type 3
21
アウターライズ屈曲で生まれれるか？ Outer rise flexure
Fig. 2
–5500–5500
–4000
–2000
–6000–8000
–6000–6000
–5500
Kuril Tren
ch
Kuril Tren
chJa
pan
Tren
ch
Japa
n Tr
ench
Joban Joban SeamountsSeamounts
Japanese Japanese SeamountsSeamounts
Erimo Erimo SeamountSeamount
DaiichiDaiichi–Kashima Kashima SeamountSeamount
NosappuNosappuFracture Fracture ZoneZone
Pacific Plate
outer-rise
Japan Trench
Site A
Site B
Site ASite B
Fig.1. Bathymetricalmap of o¡shore NE Japanusing the data of Smith &Sandwell (1997)contoured at 500mintervals (upper panel)and a model for petit-spot related to the plate£exure (lower) modi¢edfromHirano et al. (2006).White and brown boxesshow the area of Fig. 2and the region of cross-sectional model shown bythe lower ¢gure,respectively.White heavylines show theCretaceous fracture zone(Nakanishi &Winterer,1998). Area of outer-riseis shown by hatched area,which is approximatelycorrespond to shallowerarea than 6000 metersbelow sea level (mbsl) for130^140Ma ocean £oor atthe western side offracture zone and than5900mbsl for120^125Maocean £oor at the easternside of the fracture zone,respectively.
r 2008 The AuthorsJournal Compilation r Blackwell Publishing Ltd, European Association of Geoscientists & Engineers and International Association of Sedimentologists544
N. Hirano et al.
velocity region (a zone potentially representinga mantle plume) at a 410-km depth in this areawas reported by Obayashi et al. (17); however,there is no evidence for a conduit or connectionof any type between this low-velocity regionand the shallower mantle. Our geochemicalevidence strongly supports a depleted mantle(nonplume-like) source. Furthermore, the vol-ume of magma produced at the petit spotvolcanoes must have been several orders ofmagnitude less than those typical of hot spotvolcanoes. We found 4.2 to 8.5 million-year-old volcanoes at site A, suggesting episodiceruption of magma over a distance of 400 kmof plate motion. Accordingly, the petit spotvolcanic province is characterized by severalmillion years of small-volume magma produc-tion over a large area.
Temperatures of the asthenosphere beneaththe Pacific Plate have been mapped precisely
by means of surface wave tomography, show-ing that the thermal state of the asthenosphereis notably homogeneous throughout the Pa-cific Ocean (18). The temperature at a depthof 150 km is estimated to be between 1450-and 1480-C, which implies that the tempera-ture just below the Pacific Plate is 50- to 150-Clower than the solidus of dry mantle mate-rials (19). However, at this depth and temper-ature in the asthenosphere, a small percentmelt should exist in the presence of smallamounts (G1%) of H2O or CO2 (20), whichlowers the melting temperature. The highlyvesicular petit spot lavas probably representincipient partial melts that formed in the as-thenosphere in the presence of volatiles, mostplausibly CO2. If the magmas were suppliedfrom the normal asthenosphere, with emplace-ment channels controlled by tectonic fractur-ing of the overlying lithospheric plate, the low
volumes of magma output over a large areacould easily be explained.
An old and cold lithospheric plate behaveselastically and may be flexed because ofloading by an ocean island or seamount or bysubduction-related plate flexure (21). In the areabetween sites A and B, the bathymetric high (orouter rise) is aligned parallel to the Japan Trench.The Pacific Plate flexes convexly here, as itsubducts beneath Japan and yields a positivegravity anomaly (22). This flexed region iselevated 9800 m above the normal ocean floor(È6000 m below sea level at site B). Largecurvatures imposed on the preflexed lithospheremight instigate brittle fracturing (that is,bending-induced faults) (23). As for volcaniccones at site A, the volcanic features, aligned ina west-northwest to east-southeast direction, areessentially perpendicular to hinge lines on thebending plate (Fig. 1, A to C). Accordingly, itappears that magmas are brought to the surfacealong fractures parallel to the direction of themaximum horizontal compression. However,the surface compression is actually caused byextensional stresses on the base of the down-warping Pacific Plate (Fig. 3C).
Post–erosional-stage lavas on some of theHawaiian Islands, as well as submarine lavason the flexural Hawaiian Arch (because ofloading on the plate), have chemical composi-tions and tectonic emplacement mechanisms(24) that are similar to those of the petit spotlavas. It has also been proposed that alkaliclavas in the western Samoan Islands (located onthe opposite side of the main Samoan hot spotshield volcanoes) may be derived from theasthenosphere during tectonic faulting (25). Iftrue, a similar style of volcanism may be present
Fig. 2. Plot of Ne versus Ar isotopes. (20Ne/21Ne)*is equal to [(20Ne/22Ne) – (20Ne/22NeAir)]/[(21Ne/22Ne) – (21Ne/22NeAir)], showing the slopeof a mixing line on the conventional neon three-isotope plot (20Ne/22Ne versus 21Ne/22Ne) (8).Data with 20Ne/22Ne ratios of less than 10 werenot plotted (8). Literature and data for MORB andOIB may be found in the following online data-bases: Petrological Database of the Ocean Floor(www.petdb.org/) and Geochemistry of Rocks of theOceans and Continents (http://georoc.mpch-mainz.gwdg.de/georoc/).
Fig. 1. (A) Map of the northwestern PacificOcean (22) with surveyed areas noted in theblack boxes. A, site A; B, site B; C, site C. Theouter rise (G5500 m below sea level) is shaded.Bathymetric and side-scan sonar maps for site A(B and C) and site B (D and E). 50-m and 20-mcontours are indicated in (B) and (D), respectively.In (B) and (C), white rectangles indicate dive anddredge sites; dashed yellow lines indicate volcanodistributions; and the red line indicates the crosssection described in Fig. 3A. Black arrows in (D)and (E) depict the three dives with the twodredges.
REPORTS
www.sciencemag.org SCIENCE VOL 313 8 SEPTEMBER 2006 1427
(Hirano et al., 2001; 2006; 2008)
22
巨大火成岩岩石区 Large Igneous Provinces LIPs
Oceanography Vol. 19, No. 4, Dec. 2006152
as ~ 40 km, determined from seismic
and gravity studies of Iceland (Fig-
ure 1) (Darbyshire et al., 2000). Nearly
all of our knowledge of LIPs, however,
is derived from the most accessible la-
vas forming the uppermost portions of
crust. This extrusive upper crust may
exceed 10 km in thickness (e.g., Cof-
fi n and Eldholm, 1994). On the basis of
geophysical, predominantly seismic, data
from LIPs, and from comparisons with
normal oceanic crust, it is believed that
the extrusive upper crust is underlain by
an intrusive middle crust, which in turn
is underlain by lower crust (“lower crust-
al body”) characterized by compressional
seismic wave velocities of 7.0–7.6 km s-1
(Figure 2). Dikes and sills are probably
common in the upper and middle crust.
Seismic wave velocities suggest that the
middle crust is most likely gabbroic, and
that the lower crust is mafi c to ultra-
mafi c, and perhaps metamorphic (e.g.,
Eldholm and Coffi n, 2000).
Low-velocity zones have been ob-
served by seismic imaging of the mantle
beneath the oceanic Ontong Java Plateau
(e.g., Richardson et al., 2000), as well
as under the continental Deccan Traps
(Kennett and Widiyantoro, 1999) and
Paraná fl ood basalts (Vandecar et al.,
1995) (Figure 1). Interpreted as litho-
spheric roots or keels, the zones can
extend to at least 500–600 km into the
mantle. In contrast to high-velocity roots
beneath most continental areas, and the
absence of lithospheric keels in most
oceanic areas, the low-velocity zones
beneath these three LIPs apparently re-
fl ect residual effects, primarily chemi-
cal and perhaps secondarily thermal,
of mantle upwelling (Gomer and Okal,
2003). If proven to be common, roots
extending well into the mantle beneath
oceanic LIPs would suggest that such
LIPs contribute to continental initiation
as well as continental growth. Instead
of being subducted like normal oce-
anic lithosphere, LIPS may be accreted
to the edges of continents, for example,
obduction of Ontong Java Plateau ba-
salts onto the Solomon Island arc (e.g.,
Hughes and Turner, 1977), and terrane
accretion of Wrangellia (e.g., Richards
et al., 1991) and Caribbean (e.g., Kerr
et al., 1997) LIPs to North and South
America, respectively.
�² �� ²
�� ²
�²
��²
��²
��²��� ² ��� ²
-!$!'!3#!2
3)"%2)!
%-%)3(!.#%.42!,!4,!.4)#
+!2//
&%22!2
()+52!.')
3(!43+9
0!2!.! %4%.$%+!
3/54(!4,!.4)#
3/54(!4,!.4)#
-!.)()+)
/.4/.'*!6!
.!525
#!2)""%!.
"2/+%.
$%##!.
./24(!4,!.4)#
9%-%.
%4()/0)!
#/,5-")!2)6%2
%!34-!2)!.!
'!,!0!'/3
(!7!))
%-0%2/2
)#%,!.$
+%2'5%,%.
,/5)36),,%
-!3#!2%.%
.).%49%!34
7!,6)3
0)'!&%44!
#(!'/3 ,!##!$)6%
-!,$)6%
-!'%,,!.
(%33
Figure 1. Phanerozoic global LIP distribution, with transient (“plume head”) and persistent (“plume tail”) LIPs indi-cated in red and blue, respectively. LIPs are better preserved in the oceans where they are not subject to terrestrial erosional processes, off ering a prime target for scientifi c ocean drilling. Modifi ed from Coffi n and Eldholm (1994).
23
陸上の洪水玄武岩：デカン高原 flood basalt on land Deccan Trap
height ~2km, area ~500,000km2, volume 512,000 km3 68-60 Ma , multiple flood basalt
24
巨大海台：オントンジャワ海台 Oceanic Plateau: Ontong Java Plateau
largest flood basalt 4,270,000km2 119-125Ma formation 90Ma magmatic pulse
[6] On the OJP, intrabasement reflections havebeen observed along its northwestern margin[Hagen et al., 1993], along its southern margin[Phinney et al., 1999, 2004], and on its crest andeastern flank (this paper). What causes the OJP’sintrabasement reflections remains enigmatic, how-ever, because unlike lava flows drilled on volcanicpassive margins and the Kerguelen Plateau, alllavas drilled on the OJP and sampled in theSolomon Islands were erupted in a submarineenvironment [e.g., Michael, 1999; Ingle andCoffin, 2004; Roberge et al., 2005]. Furthermore,no drilling has penetrated these reflections, withthe deepest in situ penetration of OJP basementbeing 217 m at ODP Site 1185 [Mahoney et al.,
2001], and correlation between seismic data andfield observations of obducted OJP sections in theSolomon Islands is difficult because of limitedexposure of the sections.
[7] Early Cretaceous basalts on Malaita Island,termed the Malaita Volcanic Group (MVG), com-prise a monotonous sequence of pillow and mas-sive lavas and sills with rare interbedded sedimentand sedimentary rock (chert and a quite thinmudstone both calcareous and noncalcareous),and minor microgabbro, gabbro, and dolerite dikes[Petterson et al., 1997, 1999; Petterson, 2004].Rates of lava effusion for obducted OJP basalts onMalaita Island were high to very high [Petterson etal., 1997, 1999; Petterson, 2004]. The 39Ar/40Ar
Figure 1. Predicted bathymetry [after Smith and Sandwell, 1997] of the Ontong Java Plateau, outlined in black[Mahoney et al., 2001], and surrounding region. Three research cruises have acquired MCS data on the plateau: R/VKairei KR05-01 in 2005 (thick black lines), R/V Hakuho Maru KH98-1 Leg 2 (thin white lines) in 1998, and R/VMaurice Ewing EW95-11 (thin purple lines) in 1995. Deep Sea Drilling Project (DSDP) and Ocean Drilling Program(ODP) sites on the Ontong Java Plateau and in the Nauru Basin are indicated by squares; those penetrating igneousbasement are shown in red. Malaita (Solomon Islands) is indicated in red; other land area appears in gray. Blackcircles indicate the start and end of composite MCS profiles (Figure 4). Red lines indicate data shown in Figures 5–7.Inset shows zoom-in of intersection of MCS lines KR05-01 Split and KH98-01 Leg 2 501 (see Figure 2 for velocityanalysis at the intersection and Figure 7 for MCS data at the intersection).
GeochemistryGeophysicsGeosystems G3G3 inoue et al.: ontong java plateau construction 10.1029/2007GC001780inoue et al.: ontong java plateau construction 10.1029/2007GC001780
3 of 19
Inoue et al.(2008)
cally imaged mantle root. Melt would over¢ll thecrater due to thermal expansion, and it wouldalso propagate radially from the crater along frac-tures created in the brittle, surrounding litho-sphere to erupt in the proximal ocean basins(Fig. 4c). Solid mantle, not involved in the melt-ing event, would then both rise buoyantly frombeneath and £ow from the sides to replace thevolume vacated by the erupting, melted mantle(Fig. 4d). Because the melt and underlying solidmantle would ascend adiabatically, with buoyancy
generated solely by thermal expansion of ambientmantle (as opposed to excess temperatures and re-sultant dynamic buoyancy characterizing plumes),relatively minor uplift and subsidence would beassociated with emplacement of the greater OJP,and it would be in isostatic equilibrium with sur-rounding lithosphere [5,18,69] (Fig. 1b).
A large meteorite impact could result in theformation of a temporary magma lake. Such aprocess may have been common on Earth duringheavy meteorite bombardment in Early Archean
Fig. 4. Conceptual model showing possible sequence of events if an V20 km diameter bolide instigated the creation of the OJP.(a) t1 Moment of impact, water column is vaporized, 20 Myr old oceanic lithosphere (pink layer) at impact site is obliterated,uppermost asthenosphere is penetrated, and surrounding lithosphere fractures [62]. (b) t2 Moment of maximum penetration, thecrater is completely formed and melting region becomes focused. (c) t3 In¢ll of void from bottom and sides, melt also migratesout along fractures in lithosphere, refractory surrounding mantle ¢lls space vacated by out£owing magmas. (d) V120 Ma, theOJP at end of emplacement (pink layer represents V35 km thick crust). (e) V90 Ma, tectonism causing new pressure releasemelting. Scale is maintained throughout the diagram.
EPSL 6915 9-1-04 Cyaan Magenta Geel Zwart
S. Ingle, M.F. Co⁄n / Earth and Planetary Science Letters 218 (2004) 123^134 129
Ingle and Coffin (2003)
LIPs Origin - mantle plume - MOR+triplejunction - volcanic margin - impact
25
LIPsの活動と地球環境変動 LIPs and earth’s environment
Oceanography Vol. 19, No. 4, Dec. 2006156
et al., 2004) or in subaerial permafrost
settings (Figure 5). A key factor affect-
ing the magnitude of volatile release has
been whether eruptions were subaerial
or submarine; hydrostatic pressure in-
hibits vesiculation and degassing of rela-
tively soluble volatile components (H2O,
S, Cl, F) during deep-water submarine
eruptions, although low-solubility com-
ponents (CO2, noble gases) are mostly
degassed even at abyssal depths (e.g.,
Moore and Schilling, 1973; Dixon and
Stolper, 1995).
Another important factor determin-
ing the environmental impact of LIP
volcanism has been the latitude at which
the LIP formed. In most basaltic erup-
tions, released volatiles remain in the
troposphere. However, at high latitudes
(e.g., Kerguelen Plateau), the tropopause
is relatively low, allowing larger mass fl ux
(via basaltic fi ssure eruption plumes for
transport) of SO2 and other volatiles into
the stratosphere. Sulfuric acid aerosol
particles that form in the stratosphere
after such eruptions have a longer resi-
dence time and greater global dispersal
than if the SO2 remains in the tropo-
sphere; therefore, they have greater ef-
fects on climate and atmospheric chem-
istry (e.g., Devine et al., 1984; Stothers
et al., 1986). Subaerial eruptions of large
volumes of basaltic magma at high-lati-
tude LIPs over relatively brief geological
intervals, including phreatomagmatic
eruptions (an explosive eruption result-
ing from the interaction of magma with
water) (e.g., Ross et al., 2005), would
have increased potential to contribute to
global environmental effects.
Highly explosive felsic eruptions, such
as those documented from the North
and South Atlantic volcanic passive mar-
gins, the Red Sea, the Kerguelen Plateau,
and continental fl ood basalts (e.g., Bryan
Continental Flood Basalt Oceanic Plateau (off-axis)
Oceanic Crust
Stratosphere
Troposphere
SO2, HCl, Ash, [CH4]
RainoutH2O, HCl, Ash
Nucleation and Particle Growth
Increased Planetary Albedo Heterogeneous Chemistry
N2O5
ClONO2
HCl
HNO3
ClO
WarmingSO2 H2SO4
Cirrus Modification
RemovalProcesses
Infrared
hN & OH
Hydrothermal ActivitySea Level
UpwellingGatewaysLand Bridges
ContinentalCrust
Lower Crustal Body
Lower Crustal Body
Extrusives
Extrusives
Figure 5. Complex chemical and physical environmental eff ects associated with LIP formation. LIP eruptions can perturb the Earth-ocean-atmosphere system signifi cantly. Note that many oceanic plateaus form at least in part subaerially. Energy from solar radiation is hv, where h = Planck’s constant, and v = frequency of electromagnetic wave of solar radiation. Modifi ed from Coffi n and Eldholm (1994).
26
Latest pulse of Earth: Evidence for a mid-Cretaceous superplume
R. L. Larson Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island 02882
ABSTRACT ' A calculation of Earth's ocean crustal budget for the past 150 m.y. reveals a 50% to 75%
increase in ocean crust formation rate between 120 and 80 Ma. This "pulse" in ocean crust production is seen both in spreading-rate increases from ocean ridges and in the age distribu-tion of oceanic plateaus. It is primarily a Pacific Ocean phenomenon with an abrupt onset, and peak production rates occurred between 120 and 100 Ma. The pulse decreased in intensity from 100 to 80 Ma, and at 80 Ma rates dropped significantly. There was a continued decrease from 80 to 30 Ma with a secondary peak near the Cretaceous/Tertiary boundary at 65 Ma. For the past 30 m.y., ocean crust has formed at a nearly steady rate. Because the pulse is seen primarily in Pacific oceanic plateau and ridge production, and coincides with the long Creta-ceous interval of normal magnetic polarity, I interpret it as a "superplume" that originated at about 125 Ma near the core/mantle boundary, rose by convection through the entire mantle, and erupted beneath the mid-Cretaceous Pacific basin. The present-day South Pacific "super-swell" under Tahiti is probably the nearly exhausted remnant of the original upwelling. How this superplume stopped magnetic field reversals for 41 m.y. is a matter of speculation, but it probably involved significant alteration of the temperature structure at the core/mantle boundary and the convective behavior of the outer core.
35 "I
PACIFIC RIDGES (EXCEPT TETHYS)
1 10" I—
! G O N D W A N A l , R I D G E S L L
- D E C C A N , •, , - T , | i - r - -r | i • | - —i—1 t—' r >—r • i -i '-i- r i ' i i i O o < gj 7 H t: CO s D n O H -J w s o M £ £ U i - i- c/5 —f V < —r—'—r 03 > i CO
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
M I L L I O N S O F Y E A R S
Figure 1. World oceanic crust production tor past 1S0 m.y. partitioned into Pacific ridges, Gond-wana ridges (Atlantic and Indian oceans), and oceanic plateaus that sum to world total. Deccan Traps volume shown for comparison with continental flood basalts. Each point on these noncu-mulative histograms represents volume of ocean crust produced in 1 m.y. "No magnetic rever-sals" represents long Cretaceous interval of normal polarity enclosed by magnetic anomalies 34 to M0. Geologic time scale from Harland et al. (1990).
GEOLOGY, v. 19, p. 547-550, June 1991
INTRODUCTION AND METHODS Earth is a huge heat engine, fueled mainly by
the decay of the radioactive isotopes of potas-sium, uranium, and thorium and by release of the heat of crystallization at the inner/outer core boundary. The heat from this engine is dissi-pated mainly during the formation of oceanic crust in the world's ocean basins. It is generally assumed that this heat energy is produced at a nearly constant rate. However, the constancy of heat-energy dissipation has been a source of speculation for decades (Holmes, 1965), and an episodically "pulsating" Earth could account for mountain-building episodes, climatic extremes, eustatic sea-level fluctuations, and abnormal ac-cumulations of petroleum. The most recent of these pulses can now be confirmed and quanti-fied as a 50% to 75% increase in oceanic-crust production during mid-Cretaceous time. The in-itial suggestion (Larson and Pitman, 1972) of a mid-Cretaceous spreading pulse was speculative, but new evidence on magnetic lineation map-ping, magnetic reversal stratigraphic calibration, and ocean-crustal dating allow a more quantita-tive calculation of Earth's ocean-crustal budget for the past 150 m.y. The sources of error in this calculation are still finite but are substantially reduced by recent studies. One source of error, the biostratigraphic calibration of the mid- to Late Cretaceous magnetic reversal time scale, is virtually eliminated. The magnetic reversal stra-tigraphy of the Late Cretaceous at Gubbio, Italy (Alvarez et al., 1977), places the upper end of the long Cretaceous interval of normal magnetic polarity (magnetic anomaly 34) at the Santon-ian/Campanian boundary, whereas the lower end (magnetic anomaly M0) is placed with equal certainty in the early Aptian in other Ital-ian sections (Lowrie et al., 1980), and by Deep Sea Drilling Project (DSDP) Hole 417D it is placed in the western North Atlantic Keathley magnetic lineation sequence (Miles and Orr, 1979).
New magnetic anomaly studies were sum-marized on world maps of basement age by Lar-son et al. (1985) and of magnetic lineations by Cande et al. (1989). Whereas many detailed areas remain for future analysis, the world-wide distribution of magnetic lineations is, in the main, well known. Working versions of these charts and various tectonic models were used by Kominz (1984, Table 1) to compile tables of ridge crest lengths vs. spreading rates for the
547
Larson, 1991
and Gondwana (Atlantic and Indian oceans) ridges, and oceanic plateaus are shown sepa-rately. The "world total" is the sum of the lower three curves, excluding Tethys. The Deccan Traps (Courtillot et al, 1987; White and McKenzie, 1989) at 65 Ma are shown for com-parison of a well-known continental flood-basalt sequence, but are not included in the world total. Clearly, a pulse of ocean crustal produc-tion appears between 120 and 80 Ma in the world total that is predominantly the result of contributions from the Pacific ridges and Pacific oceanic plateaus. Maximum spreading rates cal-culated for individual ridges during the Creta-ceous pulse are about 17 cm/yr. This is, perhaps coincidentally, about the same as the present world-maximum spreading rates observed at the Pacific-Nazca ridge. Thus, individual ridges did not spread abnormally rapidly in the mid-Cretaceous, but average rates on the Pacific ridges were clearly higher.
The onset of the mid-Cretaceous episode is sudden and is seen in all three components (Pa-cific ridges, Gondwana ridges, and oceanic pla-teaus) of the world total. The general shapes of
the histograms for "world total" and "oceanic plateau" production are very similar. The mid-Cretaceous pulse peaked soon after its onset (between 120 and 100 Ma), after which it continued with reduced intensity from 100 to 80 Ma. This decay is partially the result of end-on ridge subduction in the Pacific as opposed to spreading-rate variations that cannot be meas-ured during the long Cretaceous normal polarity interval. However, the oceanic plateau volumes also decay in the same fashion. After 80 Ma, the world total and oceanic plateau production con-tinued to decline until about 30 Ma; a secondary peak occurred near the Cretaceous/Tertiary boundary at 65 to 60 Ma. This subsidiary pulse is mainly a result of increased production rates in the Gondwana ridge system and oceanic pla-teaus (Fig. 1). In particular, it results from the fast spreading rates on the Indian Ocean ridge system associated with the breakup of Madagas-car and India that also resulted in the Chagos-Laccadives, the Madagascar Ridge, and the Deccan Traps flood basalts.
There is no evidence in Figure 1 for the hiatus in mid-Cretaceous volcanism from 95 to 80 Ma
TABLE 1. OCEANIC PLATEAU AGES AND VOLUMES
Oceanic Age Age Volume References plateau (Biostratigraphy) (Ma) (K^KmS)
Broken Ridge-1 >Turonian=Kerguelen 90-110 5.19* ODP 754 Caribbean-2 Turonian-Campanian 75-90 20.41 DSDP 146, 149, 150, Turonian-Campanian
152, 153 Caroline Seamounts Miocene-Pleistocene 1-10 5.60 Keating et al. (1984) Chagos Laccadives early Paleocene-early Eocene 50-60 14.01* ODP 707, 712, 713,715 Crozet Plateau <Anomaly 31 <65 9.45 Cande et al. (1989) Emperor Seamounts Maastrichtian-Eocene 40-70 13.42 DSDP 192, 308, 430,
431,432, 433 Hess Rise-3 late Aptian-early Cenomanian 95-115 7.78*+ DSDP 465, 466, 310 Iceland Miocene-Holocene 0-20 8.56* Moorbath et al. (1968) Kerguelen Plateau-4 Albian-eatiy Turanian 90-110 24.86* ODP 738, 748, 750 Line Islands-5 Santonian-Campanian 75-85 10.01* DSDP 165, 315,316 Madagascar Ridge Paleocene 55-65 15.94 DSDP 246, 247 Magellan Rise Jurassic/Cretaceous boundary 140-150 3.64*+ DSDP 167 Manihiki Plateau-6 Aptian 115-125 10.40*+ DSDP 317 Marcus Wake Smts.-7 Albian-Ccnomanian 90-115 30.85+ Sager and Pringle (1988) Maud Rise late Campanian-early Maast. 70-75 2.35 ODP 690 Mid Pacific Mtns.-8 Barremian-Campanian 75-130 42.94+ Hamilton (1956); DSDP
313,463, 171 Mozambique Plateau Neocomian 130-145 5.51* DSDP 249 Nazca Ridge <Anomaly 18 <40 9.17 Cande et al. (1989) Ninetyeast Ridge early Campanian-Eocene 40-85 23.74 DSDP 217, 216, 215,
214; ODP 756,757,758 Ontong Java Plat.-9 Aptian-Albian 100-125 101.35+ DSDP 288,289; ODP 803,
opn Rio Grande Rise-10 Coniacian 85-90 7.76*
oU/ DSDP 516
Shatsky Rise Anomalies MIO- M21 130-150 9.86*+ DSDP 49, 50, 306 Jurassic/Cretaccous boundary
Wallaby Plateau-11 <Australia/India breakup 110-125 1.49* Cande et al. (1989) Walvis Ridge late Campanian-Maast. 65-75 6.85 DSDP 525, 527, 528
Note: Ages arc mainly from the oldest sediment ages at the base of Deep Sea Drilling Project (DSDP) or Ocean Drilling Program (ODP) drill sites, from dredged rocks, or from tectonic associations with other features of known age (e.g., Broken Ridge=Kerguelen Plateau). Individual DSDP and ODP drill sites are not referenced in full except for the most recent sites, ODP 803 and 807 (Leg 130 Shipboard Scientific Party, 1990), but their ages can be found in the Site Chapter for each site in the Initial reports of the Deep Sea Drilling Project and in the Proceedings of the Ocean Drilling Program. Plateau volumes derived from Schubert and Sandwell (1989). Numbered plateaus are shown in Figure 2.
•Volumes reduced from their total plateau volumes by removing a 6.5 km thickness of normal oceanic crust because these plateaus formed at spreading ridges.
+Volumes doubled to account for potential twin plateaus on the Farallon plate that have been subducted.
reported by Rea and Vallier (1983). Their pro-posed hiatus was centered on the Turanian and Coniacian stages, now thought to total only 4 m.y., from 90.5 to 86.5 Ma (Harland et al., 1990). Thus, a very short hiatus may exist dur-ing these stages that is averaged out by the histo-gram intervals in Figure 1.
By 30 Ma, the world total and oceanic-plateau production rates had nearly leveled off at about the same rates seen prior to the mid-Cretaceous pulse, although it must be remem-bered that the exclusion of a potential Tethys ridge system may underestimate Early Creta-ceous and Late Jurassic rates. During the past 30 m.y., Pacific ridge output approximated Gond-wana ridge output, while world total and oce-anic plateau formation rates remained constant.
INTERPRETATION The mid-Cretaceous pulse in ocean-crust
formation is evidenced mainly in Pacific ridge and Pacific oceanic plateau production, began relatively suddenly at 120-125 Ma, and de-creased over a long period to about 70 Ma. Spreading rates during the Cretaceous lie within the present-day range, although average Pacific rates were higher. Present-day variation in spreading rates is mainly a function of the avail-ability or absence of long subducting slabs to provide driving forces for rapid spreading. Thus, it is possible that subduction-zone initiation or rearrangement in the Mesozoic Pacific is respon-sible for the pulse. However, what we know of Mesozoic subduction from the geologic record of the rim of the Pacific basin is that subduction has been generally continuous since the Jurassic, at least in the American Cordillera and in Japan. It is also unlikely that increased driving forces from subduction could explain the synchronous onset of oceanic plateau formation. Thus, I doubt that changes in subduction-zone driving forces in the Mesozoic Pacific caused the pulse. I have not included flood basalts associated with continental breakups in the compilation of oceanic plateaus. Such flood basalts might result from decompression melting during extension, whereas the oceanic plateaus of the Pacific formed far from continents and resulted from deeper mantle sources. The Pacific oceanic pla-teaus are thickened areas of oceanic crust that resulted from an abnormally high degree of melting of mantle material. This higher degree of melting requires higher temperatures (McKenzie and Bickle, 1988) that imply deeper source lev-els in the mantle. The source material must also rise quickly and approximately adiabatically to the lithosphere in order to retain most of its original heat for massive upper-mantle melting.
The pulse in both total production and oceanic-plateau building correlates closely with the long period of normal magnetic polarity in the mid-Cretaceous, and the coincident onset of both these phenomena is especially striking. As
548 GEOLOGY, June 1991
形成年代と地球の様相 Age of global LIPs
27
Oceanography Vol. 19, No. 4, Dec. 2006 157
et al., 2002) (Figure 1), have likely also
injected both particulate material and
volatiles (SO2, CO
2) directly into the
stratosphere. The total volume of felsic
volcanic rocks in LIPs is poorly known,
but such rocks may account for a small,
but important fraction of the volcanic
deposits in LIPs. Signifi cant volumes of
explosive felsic volcanism would have
further affected the global environment
(e.g., Scaillet and MacDonald, 2006).
LIP formation may have been respon-
sible for some of the most dramatic and
rapid changes in the global environment.
Between ~ 145 and ~ 50 million years
ago, the global oceans were characterized
by chemical and isotopic variations (es-
pecially in C and Sr isotope ratios, trace
metal concentrations, and biocalcifi ca-
tion), relatively high temperatures, high
relative sea level, episodic deposition of
black shales (oceanic anoxic events), high
production of hydrocarbons, mass ex-
tinctions of marine organisms, and radi-
ations of marine fl ora and fauna. Tempo-
ral correlations between the intense puls-
es of igneous activity associated with LIP
formation and environmental changes
are far too strong to be pure coincidence
(Figure 6)—far stronger, for example,
than the association of impacts and mass
extinctions. The most dramatic example
is the eruption of the Siberian fl ood ba-
salts ~ 250 million years ago (Figures 1
and 6), which coincided with the largest
extinction of plants and animals in the
geological record (e.g., Renne and Basu,
1991). Ninety percent of all species be-
came extinct at that time. On Iceland,
the 1783–1784 eruption of Laki provides
the only historical record of the type of
volcanism that constructs transient LIPs
(e.g., Thordarson and Self, 1993). Al-
though Laki produced a basaltic lava fl ow
representing only ~ 1 percent of the vol-
ume of a typical (103 km3) transient LIP
fl ow, the eruption’s environmental im-
pact resulted in the deaths of 75 percent
of Iceland’s livestock and 25 percent of its
population from starvation.
LIPS A KEY IODP INITIATIVEUnderstanding the formation of LIPs
constitutes a fi rst-order problem in
Earth science, and as such, is a major,
high-priority initiative for the Integrated
Ocean Drilling Program (IODP) (Coffi n,
McKenzie, et al., 2001). Strong evidence
exists that many LIPs manifest a form
of mantle dynamics not clearly related
to plate tectonics; the processes involved
in their formation are critically impor-
tant for understanding both mantle and
Ker
guel
en/B
roke
n, R
ajm
ahal
Ont
ong
Java
Car
ibb
ean,
Mad
agas
car
Cen
tral
Atla
ntic
Sib
eria
Em
eish
an, P
anja
l
Kar
oo a
nd F
erra
r
Par
aná
and
Ete
ndek
a
Dec
can
Nor
th A
tlant
icE
thio
pia
and
Yem
en
Col
umb
ia R
iver Figure 6. Extinction of marine genera
versus time (continuous line, blue fi eld), modifi ed from Sepkoski (1996), compared with eruption ages of LIPs (red columns). Strong temporal cor-relation between LIPs and mass extinc-tions suggests causality. Th ree of the largest mass extinctions—the Permo-Triassic, Triassic-Jurassic, and Creta-ceous-Tertiary—coincide with eruption ages of the Siberian Traps, the Central Atlantic Magmatic Province, and the Deccan Traps, respectively. Modifi ed from White and Saunders (2005). See Figure 1 for locations of LIPs.
大絶滅イベントと関連？ Extincition of marine genera & eruption of LIPs
28