- 1 -

The Izu-Bonin-Mariana Subduction FactoryRobert J. Stern, Geosciences Dept., U. Texas at Dallas, Box 830688, Richardson TX75083-0688, rjstern@utdallas.edu

The IBM arc system lies in the Western Pacific and defines the eastern margin ofthe Philippine Sea Plate (Fig. 1). Because IBM is an endmember arc system in a numberof ways, it is an outstanding natural laboratory for studying earth’s largest geodynamicsystem, the Subduction Factory. IBM is the largest intra-oceanic convergent margin,being constructed entirely within and upon oceanic lithosphere. Lithosphere producedduring the ~48 million years that this Subduction Factory has operated makes up a regionabout the size of India, but the presently active part extends for about 300 km west of theIBM trench. IBM manifests subduction of Pacific lithosphere, with the Pacific movingNW relative to IBM, at rates that vary from about 20 mm/a south of Guam to almost60mm/a near Japan [1]. The Philippine Sea Plate itself is moving rapidly northwestward.Subducted lithosphere varies in age from mid-Jurassic (~170 Ma) outboard of theMarianas to early Cretaceous (~130Ma) adjacent to Japan (Fig. 2). The combination of aretreating upper plate and an extremely old (hence dense) subducting plate results in astrongly extensional convergent plate boundary, particularly in the southern IBM wherethese effects are maximized.

Off-ridge volcanism was common during mid- and Late Cretaceous time on thePacific Plate now outboard of southern IBM. As a result, there are significant differencesin the bulk composition of subducted sedimentary columns in the north and south,although the thickness of sediments is relatively constant at about 500m [2]. This modestsediment thickness is well below the ~1km thickness required for development of anaccretionary prism [3] and is completely subducted. Plate convergence is oblique overmost of the IBM arc, approaching pure sinistral strike-slip motion in the northernMariana arc (Fig. 3). The Wadati-Benioff zone is well-defined beneath IBM, dippingabout 45° in the north and nearly vertically in the south (Fig. 4); some of the deepestseismic activity in earth – down to 700 km - is found beneath the central Mariana arc.These and other physical constraints need to be identified and incorporated in theconstruction of realistic, 4-D models of mantle and fluid flow, thermal evolution, andfluid/melt generation, migration, and storage.

Earth’s only T-T-T triple junction defines the northern end of IBM, wherecollision continues with southern Honshu at a rate of about 40mm/a. This provides anopportunity to study terrane accretion in action [4] as well as exhuming IBM arc middlecrust [5]. These exposures of arc crust are correlatable with crustal structure inferred forin situ IBM crust (Fig. 5). Not only does northern IBM provide an unparalleledopportunity to examine the formation of juvenile continental crust, societalconsiderations compel it: continuing convergence between buoyant IBM crust andHonshu presents an imposing earthquake hazard to the greater Tokyo metropolitan area,with far-reaching implications for the global economy. An analogous opportunity tostudy deeper arc crust and upper mantle lies on the wall of the 11km-deep ChallengerDeep at the south end of the IBM arc system.

Three combined forces – retreat of the Philippine Sea Plate, subduction ofunusually dense lithosphere, and oblique convergence – combine to make IBM the moststrongly extended convergent margin on the planet. This evisceration provides uniqueopportunities to monitor the Subduction Factory. Extension in the southern IBM isoriented trench-parallel in the forearc and trench-normal in the back-arc [6]. Thisstrongly extensional regime provides three opportunities - forearc, active arc, and back-arc basin - to sample Subduction Factory fluids and melts, more than any otherconvergent margin on the planet. The Mariana forearc contains the only knownoccurrences of subduction-related serpentinite mud volcanoes (SMV) [7] on this planet.Flows from Mariana SMV contain fragments of blueschist from the subducted slab [8] as

- 2 -

well as abundant mantle fragments [9]; several SMV are actively venting slab-derivedfluids, some of which support active chemosynthetic communities.

Magmatic activity is expressed differently within and along the IBM arc. A trueback-arc basin (BAB), with seafloor spreading, is developed only in the Mariana Trough[10], although well-developed inter-arc basins are developed in the Bonin Arc RiftedZone farther north (Fig. 6) [11, 12]. The active magmatic arc is largely submarine butwith abundant subaerial volcanoes [13, 14]; strong variations in magmatic compositionsaffect both submarine and subaerial edifices. BAB lavas are often aphyric or have pillowrim glass that can readily be separated, and a substantial proportion are primitive basalts;these characteristics suit BAB lavas for analysis using approaches perfected for MORB.Arc lavas are predominantly porphyritic so that bulk compositions generally do notcorrespond to magmatic liquids, and should not be studied using the techniquesappropriate for studying aphyric or glassy samples. Accumulation of plagioclasephenocrysts in particular has led to a misperception that mafic members are dominantlyhigh-Al basalts when in fact aphyric samples or glass inclusions in phenocrysts aretholeiites. Primitive compositions (Mg#>65) are uncommon among IBM arc lavas, sofractionation conditions and history need to be resolved. Old techniques for studyingpetrography should be revived and new techniques will have to be perfected if we are tounderstand the magmatic evolution of porphyritic arc lavas. In future studies,petrographic descriptions should be reported with major and trace element data. Theneed to find new ways to study evolved, porphyritic samples promises to revivetraditional petrography as well as stimulate developments in quantitativepetrography/image analysis and microbeam analytical techniques. These techniques willalso aid investigations of abundant cumulate xenoliths, found in the lavas of several IBMvolcanoes. One issue that awaits resolution is the abundance of felsic material in theIBM arc. IBM has traditionally been thought to have a basaltic bulk composition, butrecent evidence from geophysics [15], exposures in the collision zone [5], glassinclusions in phenocrysts [16], and the abundance of felsic tephra in DSDP cores [17]indicates that felsic rocks comprise an important part of the IBM arc.

In spite of the different eruptive styles and extent of fractionation for arc andBAB, there are strong compositional affinities between arc and BAB suites, whichprovide different perspectives on important controversies and enigmas. The trace elementsignatures of these lavas strongly manifest the ‘subduction component’: enrichments inlarge-ion lithophile elements and depletion in high-field strength elements, bothcompatible and incompatible (Fig. 7). All arc and most BAB lavas have elevated watercontents [12, 18] such that it is controversial the extent to which melts are generated bydecompression [19] or fluxing by hydrous fluids [20]. One abiding mystery concerns howwater gets into the source of Mariana BAB magmas when the subducted slab does not liebeneath the spreading ridge? In contrast to widespread recognition that water in IBMmelts is recycled from subducted materials, the source of other elements in IBM melts isless clear. In particular, controversy continues regarding the extent to which the IBM‘subduction component’ manifests fractionations imposed when elements are transferredfrom the subducted plate to the overlying mantle wedge as opposed to being developedduring re-equilibration of hydrous fluids and melts with convecting mantle. U-Thdisequilibria studies indicate that strong fractionation of radionuclides occurred withinthe last 30 kyr [21], providing timescales for fluid-mediated fractionation but where andhow this occurs are unresolved. There is abundant evidence that some subductedcomponents are recycled and can be found in young lavas, especially elevated watercontents, ∂

7Li [22], B/Be, ∂

11B [23], ∂

34S [24], and

207Pb/

204Pb [25]. These studies

differently emphasize roles of subducted sediments and altered oceanic crust. Rare gasdata is lacking for arc lavas but BAB lavas contain recycled atmospheric Ar in spite ofmantle-like

3He/

4He [26]; similar datasets for arc lavas promise to provide important

constraints. The 10

Be signal is muted probably because subducted sediment are mucholder than the half-life of this isotope [27]. Other isotopic and trace element data sets do

- 3 -

not readily allow recycled components to be identified: O, Sr, Nd, and Pb isotopiccompositions as well as K/Rb and K/Ba are remarkably constant in arc lavas despitebeing extremely heterogeneous in subducted components. This homogeneity requiresefficient mixing of subducted components or effective re-equilibration of ascendingfluids with convecting mantle. It is critical to resolve the nature of the mantle sourcebeneath IBM; there are strong indications from Nd, Hf, Pb, and Os data that this mantlehas affinities to that beneath the Indian Ocean [28-30].

Disparate geochemical and isotopic data sets and controversial conclusionsprovide important constraints for understanding how the IBM Subduction Factoryoperates. It will be an exciting challenge for scientists with different talents andperspectives to synthesize these observations and use these to develop and testhypotheses to better understand the operation and budget of this outstanding example ofthe Subduction Factory. A critical aspet will be to collect and distribute representativesamples to analysts and to communicate these results to modelers, and for modellers totell analysts what kinds of information they need. Another critical effort will be theseismic imaging of the IBM Subduction Factory, with as high a resolution as possible.

1. Seno, T., S. Stein, and A.E. Gripp, A model for the motion of the Philippine Sea Plateconsistent with NUVEL-1 and geological data. Journal of Geophysical Research,1993. 98(B10): p. 17,941-17,948.

2. Plank, T. and C. Langmuir, The chemical composition of subducting sediment and itsconsequence for the crust and mantle. Chemical Geology, 1998. 145: p. 325-394.

3. Le Pichon, X., P. Henry, and S. Lallemant, Accretion and erosion in subductionzones: The role of fluids. Annual Reviews of Earth and Planetary Sciences, 1993. 21:p. 307-331.

4. Taira, A., S. Saito, and others, Nature and growth rate of the Northern Izu-Bonin(Ogasawara) arc crust and their implications for continental crust formation. TheIsland Arc, 1998. 7: p. 395-407.

5. Kawate, S. and M. Arima, Petrogenesis of the Tazawa plutonic complex, centralJapan: Exposed felsic middle crust of the Izu-Bonin-Mariana arc. The Island Arc,1998. 7: p. 342-358.

6. Wessel, J.K., et al., Extension in the northern Mariana inner forearc. Journal ofGeophysical Research, 1994. 99(B8): p. 15,181-15,203.

7. Fryer, P., et al., Serpentine bodies in the Forearcs of Western Pacific ConvergentMargins: Origin and Associated Fluids, in Active Margins and Marginal Basins ofthe Western Pacific Convergent Margins, B. Taylor and J. Natland, Editors. 1995,American Geophysical Union: Washington DC. p. 259-270.

8. Fryer, P., C.G. Wheat, and M. Mottle, Mariana blueschist mud volcanism;implications for conditions within the subduction zone. Geology, 1999. 27: p. 103-106.

9. Parkinson, I.J. and J.A. Pearce, Peridotites from the Izu-Bonin-Mariana forearc(ODP Leg 125); evidence for mantle melting and melt-mantle interaction in a supra-subduction zone setting. Journal of Petrology, 1998. 39: p. 1577-1618.

10. Fryer, P., Geology of the Mariana Trough, in Back-arc Basins, B. Taylor, Editor.1996.

11. Taylor, B., et al., Structural development of Sumisu Rift, Izu-Bonin Arc. Journal ofGeophysical Research, 1991. 96: p. 113-129.

12. Stolper, E. and S. Newman, The role of water in the petrogenesis of Mariana Troughmagmas. Earth and Planetary Science Letters, 1994. 121: p. 293-325.

13. Bloomer, S.H., R.J. Stern, and N.C. Smoot, Physical volcanology of the submarineMariana and Volcano Arcs. Bulletin of Volcanology, 1989. 51: p. 210-224.

- 4 -

14. Yuasa, M., et al., Submarine topography of seamounts on the volcanic front of theIzu-Ogasawara (Bonin) Arc. Bulletin of the Geological Survey of Japan, 1991. 42: p.703-743.

15. Suyehiro, K., et al., Continental crust, crustal underplating, and low-q upper mantlebeneath an oceanic island arc. Science, 1996. 272: p. 390-392.

16. Lee, J. and R.J. Stern, Glass inclusions in Mariana Arc phenocrysts; A newperspective on magmatic evolution in a typical intra-oceanic arc. Journal of Geology,1998. 106: p. 19-33.

17. Lee, J., R.J. Stern, and S.H. Bloomer, Forty million years of magmatic evolution inthe Mariana Arc: The tephra glass record. Journal of Geophysical Research, 1995.100: p. 17,671-17,687.

18. Newman, S., E. Stolper, and R.J. Stern, H2O and CO2 in magmas from the Marianaarc and back arc system. Geochemistry, Geophysics,Geosystems, 2000. 1.

19. Plank, T. and C.H. Langmuir, An evaluation of the global variations in the majorelement chemistry of arc basalts. Earth and Planetary Science Letters, 1988. 90: p.349-370.

20. Eiler, J.M., et al., Oxygen Isotope Geochemistry of Oceanic-Arc lavas. Journal ofPetrology, 2000. 41: p. 229-256.

21. Elliott, T., et al., Element transport from slab to volcanic front at the Mariana arc.Journal of Geophysical Research, 1997. B102: p. 14,991-15,019.

22. Moriguti, T. and E. Nakamura, Across-arc variation of Li isotopes in lavas andimplications for crust/mantle recycling at subduction zones. Earth and PlanetaryScience Letters, 1998. 164: p. 167-174.

23. Ishikawa, T. and F. Tera, Two isotopically distinct fluid components involved in theMariana Arc; evidence from Nb/B ratios and B, Sr, Nd, and Pb isotope systematics.Geology27, 1999: p. 83-86.

24. Alt, J.C., W.C. Shanks III, and M.C. Jackson, Cycling of sulfur in subduction zones;the geochemistry of sulfur in the Mariana island arc and back-arc trough. Earth andPlanetary Science Letters, 1993. 119: p. 477-494.

25. Woodhead, J.D. and D.G. Fraser, Pb, Sr, and 10Be isotopic studies of volcanic rocksfrom the Northern Mariana Islands. Implications for magma genesis and crustalrecycling in the Western Pacific. Geochimica et Cosmochimica Acta, 1985. 49: p.1925-1930.

26. Ikeda, Y., et al., Noble gases in pillow basalt glasses from the northern MarianaTrough back-arc basin. The Island Arc, 1998. 7: p. 471-478.

27. Tera, F., et al., Sediment incorporation in island-arc magmas: Inferences from 10Be.Geochimica et Cosmochimica Acta, 1986. 50: p. 535-550.

28. Hickey-Vargas, R., J.M. Hergt, and P. Spadea, Ocean-type isotopic signature inWestern Pacific marginal basins: Origin and significance, in Active Margins andMarginal Basins of the Western Pacific, B. Taylor and J. Natland, Editors. 1995,American Geophysical Union: Washinton DC. p. 175-197.

29. Parkinson, I.J., C.J. Hawkesworth, and A.S. Cohen, Ancient mantle in a modern arc:Osmium isotopes in Izu-Bonin-Mariana forearc peridotites. Science, 1998. 281: p.2011-2013.

30. Pearce, J.A., et al., Hf-Nd Element and Isotope Perspective on the Nature andProvenance of Mantle and Subduction Components in Western Pacific Arc-Basinsystems. Journal of Petrology, 1999. 40: p. 1579-1611.

31. Nakanishi, M., K. Tamaki, and K. Kobayashi, Magnetic anomaly lineations from LateJurassic to Early Cretaceous in the west-central Pacific Ocean. Geophysical JournalInternational, 1992. 109: p. 701-719.

- 5 -

32. McCaffrey, R., Estimates of modern arc-parallel strain rates in fore arcs. Geology,1996. 24: p. 27-30.

33. Chiu, J.-M., B.L. Isaaks, and R.K. Cardwell, 3-D configuration of subductedlithosphere in the western Pacific. Geophysical Journal International, 1991. 106: p.99-111.

34. Peate, D.W. and J.A. Pearce, Causes of spatial compositional variations in Marianaarc lavas: Trace element evidence. The Island Arc, 1998. 7: p. 479-495.

- 6 -

Figure Captions:

Figure 1: IBM arc system, showing extent of crust generated over the 48 Ma life of thearc (fossil plus active) as well as those components which comprise the presently activeIBM Subduction Factory. Also shown is the location of the deepest place on the face ofthe earth, the Challenger Deep (~11km deep) and the IBM collision zone.

Fig. 2: Seafloor feeding into the IBM Subduction Factory, modified after [31]. Arrowsare relative velocities of the Pacific Plate with respect to the Philippine Sea Plate, inmm/a, after [1]. DSDP and ODP sites sampling units being subducted beneath IBM areshown as well. Note that the sedimentary section being subducted beneath northern IBMhas fewer volcanics and volcaniclastics than that being subducted beneath the southernIBM.

Fig. 3: Obliquity of convergence between the Pacific and Philippine Sea plates, asinferred from earthquake slip vectors and modified after [32]. Note that convergence ishighly oblique over much of the IBM arc system.

Fig. 4: Generalized topology of IBM Wadati-Benioff Zone, modified after [33]. Twoperspectives are shown, with contours colored at every 100 km depth.

Figure 5: Structure of IBM arc crust at 32°15’N, modified after [15]. Verticalexaggeration is about 10x. Note the crustal thickness of 22km is nearly four times that ofoceanic crust but only about half that of normal continental crust.

Figure 6: Along-strike profiles of the IBM arc system, from Japan (left) to Guam (right).The thick solid line shows the bathymetry and topography along the volcanic axis of theactive arc, with the thin dashed horizontal line marking sea level. The approximatelocations of the principal island groups (Izu, Bonin-Volcano, and Mariana) are shown.Submarine volcanoes (and the Sofugan Tectonic Line, STL) are given as italicizedabbreviations: Ku, Kurose;Ms, Myojin-sho; Do, Doyo; Kk, Kaikata; Kt, Kaitoku;F,Fukutoku-oka-no-ba; HC, Hiyoshi Volcanic Complex, Nk, Nikko; Fj, Fukujin, Ch,Chamorro, D, Diamante; R, Ruby, E, Esmeralda; T; Tracy. Subaerial volcanoes are givenas normal abbreviations: O, Oshima; My, Miyakejima; Mi, Mikurajima; H, Hachijojima;A, Aogashima; Su, Sumisujima, T, Torishima; Sg, Sofugan; Nishinoshima; KIJ, Kita IwoJima; IJ, Iwo Jima; MIJ, Minami Iwo Jima; U, Uracas; M, Maug; As, Asuncion; Ag,Agrigan; P, Pagan; Al, Alamagan; G; Guguan; S, Sarigan; An, Anatahan. Dominantcompositions of arc segments are also indicated. Locations of important zones of intra-arc and back-arc extension in the north (Bonin Arc Rifted Zone) and south (MarianaTrough Back-Arc Basin) are marked. The thick dashed line shows the maximum depthin the trench along its strike. Frontal arc elements are not shown, but consist of the Boninor Ogasawara Islands between 26° and 28°N and the Mariana frontal arc islands between13° and 16°N. ICZ = IBM collision zone.

Fig. 7: ‘Spider’ diagram for Mariana arc lavas. Elements are listed in order of increasingcompatibility in mantle minerals; data for typical Maraian arc lavas is from [34]. Noticestrong enrichments in LIL and depletions in HFSC, including Nb and Ta.

Japan

*Tokyo

..

Pala

uK

yush

uR

idge

-

ShikokuBasin

Parece VelaBasin

... .......

. ...

...

........

.

Belau

.

30¡

40¡

20¡

10¡

140¡

WestPhilippine

Basin

Active Spreading Ridge(Mariana Trough)

Extinct Spreading Ridge

Active IBM system Fossil IBM system

IBMARC

SYSTEMPacificPlate

Izu Arc

Bonin Arc

Mariana A

rc

Challenger DeepYap

Guam

IBMCollision

Zone

Jurassic

Ontong JavaPlateau

Hawaii-Emperor Smts.

Mid-PacificMountains

Cretaceous

Age of Western Pacific Seafloor

Cretaceous

Jurassic

**

*

*

v v vv v v

vv v v v

v v v v vv v v v v

v

vvvvvv

v v

vvv

v v

v

v v v v v

vvvv

vv

v

v

v v v v

vvv

vv vI-B

-M 801800

452

802

198307197 196

195 194

462

Oligocene and younger crust

Oceanic Plateaus (mostly Cretaceous)

vv Mid-Cretaceous Flood Basalts

* DSDP & ODP sitesTrenches

*

Caroline Ridge

1149

Mid-PacificMountains

ShatskyRise

*

***

**

140°E 160 180°

*

v

v

Mariana Trench21

40

57

v

plate convergence obliquity(from earthquake slip vectors) arc-parallel fore-arc slip rate

North South

Izu Bonin Mariana

0 1000 2000 3000 km

34.1° 25.2° 16.9°N 142°E

60

30

0

-30

-60

-90

orthogonal convergence

sinistralstrike-slip

Obl

iqui

ty (

°) o

r ra

te (

mm

/y)

N

N

500 km

128°E 138° 148°

28°N

18°

8°

Izu Trench

Mariana Trench

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

25

20

15

10

5

0

Dep

th (

km)

138 139 140 141 142

Longitude ( ¡E)

2

143

Volcanic Arc

Trench

Mantle Crust Moho

P-w

ave

Sei

smic

Vel

oci

ty (

km/s

ec)

Mantle

Crust

Fore-arc

Moho

Pacific Plate

100 km

262832 3034

KIJ+1SL

-2

-4

24 22 20 18 16

IJ HCFj As

Ch Ag P

Al G S DO

My Mi H A

Ms Su T Sg N

AnMIJ

-6

-8

-10

E T

14

(Km)

Degrees N latitude

R

500 km

Trench

Active Arc

U

Izu Islands Bonin &Volcano Islands Mariana Islands

Nk

Mariana Trough Back-Arc Basin Bonin Arc Rifted Zone

FSTLDo

Ku MICZ

Medium-K Basalt Shoshonite Low-K Tholeiite

KkKt

Cs Ba U Ta Pb Sr HfHREE

HREETiLa Nd EuThRb K Nb Ce Pr P Zr Sm

100

10

1Sam

ple/

MO

RB Mariana Shoshonites

Mariana Islands

Mariana Trough