ELEMENTS, VOL. 10, PP. 115–120 APRIL 2014115

1811-5209/14/0010-115$2.50 DOI: 10.2113/gselements.10.2.115

Izu-Bonin-Mariana Forearc Crust as a Modern Ophiolite Analogue

SUBDUCTION INITIATION AND FOREARCSThe initiation of subduction and the subsequent birth and early development of island arcs are still poorly under-stood (e.g. Stern 2004). Exposed “naked forearcs,” where neither accretion of material from the subducting plate nor volcanism or sedimentation after the earliest stage of arc formation occurred to cover the basement litholo-gies, are particularly valuable for understanding how subduction zones start (Stern et al. 2012). The Izu-Bonin-Mariana (IBM) forearc (FIG. 1) is an archetypal example with widespread exposures of early-arc lavas, and it is thus an excellent location for investigating subduction initia-tion and subsequent island arc evolution. Understanding the igneous stratigraphy of the IBM forearc has been slow because most outcrops of intraoceanic forearcs are deeply submerged and are diffi cult to study directly. In recent years, we have examined the forearc crustal section of the IBM along the trench slope via submersible observa-tions and extensive sampling (Ishizuka et al. 2006, 2011a; Reagan et al. 2010). We have discovered that the basement lithologies here represent magmatism associated with subduction initiation and that this magmatism occurred nearly synchronously along a zone that is up to 300 km wide and 3000 km long (Stern and Bloomer 1992; Bloomer et al. 1995).

FOREARC STRATIGRAPHY OF THE IZU-BONIN-MARIANA ARCThe Bonin Islands (FIG. 1) expose an intact record of magmatism produced early in the history of the IBM arc. These islands are known as the type locality for boninite, a high-Mg andesite lacking plagioclase phenocrysts and with exceedingly low concentrations of titanium and rare earth elements (REEs). Geochronological studies have revealed that the eruption of boninites in the IBM arc lasted for 3 to 4 million years after starting at ~48 Ma. These boninitic lavas, therefore, constitute the oldest volcanic sequence exposed on the islands (Ishizuka et al. 2006,

2011a). The Mariana section of the IBM arc also has a forearc high between the Mariana Trench and the volcanic front (FIG. 1); this high includes the islands of Guam, Rota,

Recent geological and geophysical surveys in the Izu-Bonin-Mariana forearc have revealed the occurrence on the seafl oor of oceanic crust generated in the initial stages of subduction and the earliest stage

of island arc formation. The earliest magmatism after subduction initia-tion generated forearc basalts, and subsequently, boninitic and tholeiitic to calc-alkaline lavas were produced. Collectively, these rocks make up the extrusive sequence of the Izu-Bonin-Mariana forearc oceanic crust. This volcanic stratigraphy and its time-progressive development are analogous to those documented from many suprasubduction zone ophiolites. Most supra-subduction zone ophiolites may be on-land fragments of forearc oceanic crust, produced during the initiation of subduction and the early stages of island arc development.

KEYWORDS: Izu-Bonin-Mariana arc, subduction initiation, forearc stratigraphy, suprasubduction zone ophiolite, boninite, geochronology

Osamu Ishizuka1,2, Kenichiro Tani2, and Mark K. Reagan3

1 Institute of Geology and GeoinformationGeological Survey of Japan/AIST, Central 7, 1-1-1 Higashi, Tsukuba, Ibaraki, 305-8567, JapanE-mail: o-ishizuka@aist.go.jp

2 IFREE, JAMSTEC, 2-15 Natsushima Yokosuka, Kanagawa, 237-0061, Japan

3 Department of Geoscience, 121 Trowbridge Hall University of Iowa, Iowa CityIA 52242, USA

Bonin forearc sheeted dykes encountered

at 5555 meters below sea level

140˚ 145˚ 150˚E10˚

15˚

20˚

25˚

30˚

35˚N

6K1093

Japan

Marianaforearc

Izu-Boninarc

Marianaarc

DSDPSites 458,459

Bonin Ridge(Bonin forearc)

Bonin Islands

6K1154

7K417

ODPSite792

Saipan

Guam

ODPSite793

OmachiSeamountMukojima

ChichijimaHahajima

-8000

-6000

-4000

-2000

0

water depth (m)

0 500km

FIGURE 1 Bathymetric map of the Izu-Bonin-Mariana arc–trench system, showing the locations of the drill sites (boxes) and dives (stars) described in this paper.

ELEMENTS APRIL 2014116

Tinian, and Saipan. The early-arc rocks here are transitional between boninite and tholeiitic arc lavas and have ages of ~44 Ma (e.g. Reagan et al. 2008). In both sections, tholeiitic to calc-alkaline lavas with relatively normal major and trace element compositions have ages of 44–45 Ma.

In addition to these subaerially exposed sections, submarine portions of the forearc highs have been investigated during the past several years (FIG. 1; Ishizuka et al. 2006, 2011a; Reagan et al. 2010). The investigations in the Izu-Bonin section started with bathymetric surveys and rock sampling by dredging, with the purpose of documenting the struc-ture and stratigraphy of the forearc crust. Direct observa-tions of the crustal section along the trenchward slope of the forearc highs using both a manned submersible (Shinkai 6500) and a remotely operated vehicle (Kaiko), which are capable of diving to depths of 6500 m and 7000 m, respec-tively, followed the initial bathymetric surveys (Ishizuka et al. 2011a). The southern Mariana forearc was also exten-sively surveyed by early dredging and drilling, followed by Shinkai 6500 dives (Reagan et al. 2010). The observed geological sections of the Izu-Bonin and Mariana forearcs are nearly identical. The stratigraphic section of the forearc crust overlying the peridotites comprises, from bottom to top (FIG. 2), (1) gabbroic rocks, (2) a sheeted dike complex (FIG. 3A), (3) basaltic lava fl ows (FIG. 3B), (4) lavas and dikes of boninite and their differentiates (FIG. 3C, D), (5) transitional high-Mg andesites, and (6) tholeiitic and calc-alkaline arc lavas.

The oldest volcanic products of the IBM arc are forearc basalts (FABs), consisting mainly of pillow lavas and hyalo-clastites (FIG. 2; Reagan et al. 2010). The lavas encountered in DSDP (Deep Sea Drilling Project) sites 458 and 459 in the central Mariana forearc show a progression from FAB-like lavas at the bottom to nearly boninitic lavas on top, linking these two lava types in time and space (Reagan et al. 2010).

The sheeted dikes occur at the same or lower stratigraphic levels relative to the FAB lavas, and they have mineral assemblages and chemical compositions identical to those of the FAB lavas. The formation of sheeted dikes in the IBM arc–trench system illustrates that magmatism occurred in an extensional stress fi eld; in addition, these dikes are key evidence for seafl oor spreading and ocean crust production in a forearc tectonic setting (e.g. Umino et al. 2008). These observations indicate that the fi rst magmatic activity in the IBM arc was associated with seafl oor spreading.

The sequence underlying the FABs in the Bonin Islands area consists of medium-grained, undeformed, orthopyroxene-bearing gabbros, predominantly composed of plagioclase, clinopyroxene, and orthopyroxene, with accessory magne-tite, apatite, and zircon. The gabbros encountered in the Mariana forearc have little or no orthopyroxene, but are otherwise similar in texture and composition to gabbros from the Bonin Islands area.

The peridotites at the lowest stratigraphic level of the IBM forearc section are dominated by harzburgite with dunite veins (Morishita et al. 2011a). Two different dunites are present: dunite bearing spinels with high atomic Cr/(Cr+Al) ratios (i.e. chromium number, Cr#, greater than 0.8), and dunite containing spinels with medium Cr# values (0.4–0.6) (Morishita et al. 2011a). These two types of dunite were in equilibrium with two distinctive melts—boninitic melt and mid-ocean ridge basalt (MORB)–like melt. This interpretation is consistent with the volcanic stratigraphy observed in the IBM forearc, where FABs and boninitic lavas progressively overlie the peridotites. From this observation, we infer that the magmas produced in the earliest phase of arc formation were derived from and interacted with the upper mantle in the forearc basement.

AGE CONSTRAINTSFABs from the Bonin forearc were dated at between 48 and 52 Ma by 40Ar/39Ar techniques (FIG. 4; Ishizuka et al. 2011a). These ages are older than the oldest age for the boninites (48.2 Ma; Ishizuka et al. 2006), confi rming that FAB predated boninite eruption. These age relationships mean that the oldest magmatic activity in the IBM arc dates back to ~52 Ma, which closely matches the timing of the change in the Pacifi c Plate motion (~50 Ma; Sharp and Clague 2006). The available U–Pb ages of zircon from Bonin forearc gabbros are 51.6 and 51.7 Ma, indicating that these intrusive rocks formed contemporaneously with the forearc basalts (FIG. 4; Ishizuka et al. 2011a). Boninites from the Bonin Islands have 40Ar/39Ar ages of 46.0–48.2 Ma, and the overlying transitional high-Mg andesites have ages of 44.3–44.74 Ma (FIG. 4). The uppermost sections consist of tholeiitic and calc-alkaline arc lavas, and these are younger than 44 Ma (Ishizuka et al. 2006).

FABs from the Mariana forearc also yielded a 40Ar/39Ar age of 51.1 Ma (Ishizuka et al. 2011a), indicating that volcanism associated with subduction initiation occurred contemporaneously along the entire length of the IBM arc. Lavas recovered from DSDP sites 458 and 459, with transitional compositions between FAB and boninite, were dated at 49 Ma (Cosca et al. 1998), and lavas transitional between boninite and tholeiitic arc lavas on Guam were dated at about 44 Ma (Meijer et al. 1983). The eruption of high-silica rhyolite on Saipan (45.1 Ma; Reagan et al. 2008) marked the fi rst appearance of arc volcanism with relatively normal trace element compositions. Tholeiitic to calc-alkaline arc lavas on Rota, Saipan, and Guam erupted

Troodos Semail

gabbro/plagiogranite

sheeted dikes

gabbro/plagiogranite

Geotimes

Alleyunit

IBM forearc

gabbro

sheeteddikes

forearcbasalts

peridotite

arc tholeiites and calc-alkalinerocks

high-Mgandesite

boniniteandesiterhyolite

sheeted dikes

Lasail

Upper PillowLavas

Lower PillowLavas

(boniniteandesite-rhyolite)

(boninite, basalt,andesite)

ultramaficrocks

ultramaficrocks

FIGURE 2 Schematic stratigraphic section of the Izu-Bonin-Mariana (IBM) forearc lithosphere compared with

sections of the Troodos (Cyprus) and Semail (Oman) ophiolites. MODIFIED AFTER ISHIZUKA ET AL. (IN PRESS)

ELEMENTS APRIL 2014117

after 42 Ma (FIG. 4; Reagan et al. 2008). These age relation-ships suggest that the crustal section exposed between the trench and the forearc high along most or all of the IBM system was generated over approximately 7–8 million years while the area transitioned from mantle upwelling and near-trench extension immediately after subduction initia-tion to normal mantle counterfl ow.

MAGMATIC EVOLUTIONThe geochemical characteristics of FABs are compatible with their being generated from clinopyroxene-impover-ished residual mantle. This mantle was left over from an earlier stage of decompression melting of possibly depleted asthenospheric mantle that upwelled when the Pacifi c Plate began to subduct beneath the Philippine Sea Plate (Reagan et al. 2010). FABs have light REE (LREE)–depleted patterns like MORB, but have lower concentrations of REEs and high-fi eld-strength elements (HFSEs), such as Nb, Ta, and Zr (FIG. 5A, B). For example, FABs have signifi cantly lower Ti concentrations and Ti/V ratios compared to basalts from the Philippine Sea Basin (FIG. 5C; Reagan et al. 2010; Ishizuka et al. 2011a), and they generally have lower LREE/HREE (heavy REE) ratios compared to N-MORB (normal MORB) and backarc basin basalts from the Philippine Sea Basin (e.g. Hickey-Vargas 1998; Ishizuka et al. 2009).

Most FABs lack a clear enrichment in large-ion lithophile elements (LILEs, such as Ba, Sr, Pb) compared to REEs and HFSEs, implying that, like MORB, they received little or no input from slab-derived material (FIG. 5A). Nevertheless, transitional FAB glasses recovered from DSDP sites 458 and 459 are clearly enriched in LILEs as a result of the modifi -cation of their melt source by subduction-derived fl uids (Reagan et al. 2010).

Boninite was formed by partial melting of depleted harzburgite that was hydrated by slab-derived fl uids. Heat was supplied by an upwelling asthenosphere associated with the sinking slab. Transitional and boninitic lavas share some geochemical characteristics with FABs, such as low Ti, low HFSE contents, and very low Ti/V ratios (FIG. 5A, C). These geochemical features indicate a depleted mantle source for the magmas of the boninites and transi-tional lavas. There also are some clear differences between the two magma types in terms of their melt sources. For example, boninites have a compositional signature indica-tive of slab-derived material, such as high concentrations of alkali metals, alkaline earths, and U compared to REE concentrations (FIG. 5A, B).

Boninitic rocks have been found in the same stratigraphic position at several sites along the entire ~3000 km length of the IBM forearc, suggesting that boninite volcanism occurred contemporaneously along the entire forearc, beginning at about 48 Ma (FIG. 4). The wide distribution of boninitic rocks rules out the possibility of a point heat source for the genesis of their magmas, and is consistent with a model of boninite magma genesis involving high temperatures at shallow depths along the length of the IBM arc.

Following boninite magmatism, transitional high-Mg andesite (Taylor et al. 1994; Ishizuka et al. 2006; Reagan et al. 2008) and then tholeiitic to calc-alkaline arc magmas started to erupt after ~44 Ma. This activity marked the beginning of steady-state subduction and convection in the mantle wedge. The transitional suites lavas from Guam, with compositions similar to those of the high-Mg andesite lavas from the Bonin Ridge, erupted at about 44 Ma (FIG. 4). This observation means that the eruption of high-Mg andes-ites (transitional suites) with REE and TiO2 concentrations

FIGURE 3 Photos of samples and outcrops of the IBM forearc. Locations of the dive sites are shown in FIGURE 1.

(A) Sheeted dike outcrop at 5555 meters below sea level (mbsl) (Shinkai 6500 dive 6K1154). (B) Aphyric pillow basalt in the Mariana forearc at 6015 mbsl (Shinkai 6500 dive 6K1093).

The total height of the lava outcrop is over 50 m. (C) Boninite dike swarm along the east coast of Chichijima Island, Bonin Islands. The cliff height is about 50 m. Many dikes are oriented NNW-SSE. (D) Boninite pillow lavas on the coast of Chichijima Island.

A

C

B

D

ELEMENTS APRIL 2014118

higher than those of boninites from the Bonin Ridge was also coeval along the entire length of the arc. After 44 Ma, tholeiitic to calc-alkaline arc magmas formed, and these share many geochemical characteristics with the later arc magmas. Note, however, that transitional boninites erupted at the same time as tholeiitic to calc-alkaline arc lavas erupted in different sections of the IBM arc between 45 and 43 Ma. We interpret these relationships to mean that there was an ~2-million-year time period during which mantle counterfl ow replaced the upwelling mantle regime and that this switch in the convection regimes occurred at different times in different portions of the arc.

The arc lavas that began erupting after 44 Ma have higher TiO2, Zr, Hf, and REE concentrations than the older boninitic lavas (FIG. 5A). These geochemical features imply that the mantle involved in the younger IBM arc magmatism was less depleted than the mantle sources of the magmas older than 44 Ma. This phenomenon can be explained by a change from simple upwelling to counter-fl ow of the asthenosphere in the newly established mantle wedge, causing partial melting to move deeper into the mantle (Ishizuka et al. 2006). During this process, the magmatic axis of the arc retreated from the trench, which in turn facilitated cooling of the forearc mantle by contin-uous subduction.

DISCUSSION AND LINKS TO SUPRASUBDUCTION ZONE OPHIOLITES A possible analogy between western Pacifi c forearc oceanic crust and supraduction zone (SSZ) ophiolites has been discussed for more than 20 years (e.g. Bloomer et al. 1995; Stern et al. 2012). These discussions were based on the understanding of forearc oceanic crust derived from ODP (Ocean Drilling Project) and DSDP boreholes and dredge sampling in the IBM and Tonga forearcs (Bloomer et al. 1995). New key observations obtained from the IBM forearc during the last several years include the following:

1. MORB-like tholeiitic basalts (FABs) with compositions that differ from nearby backarc lavas have been found underlying the boninitic lavas.

2. Sheeted dikes, clear evidence of seafl oor spreading, have been found.

3. Lavas transitional between FABs and boninites have been recognized, linking these two magma types in time and space.

4. Precise geochronology of forearc sections along the IBM arc has documented the timing, duration, and scale of each of the magmatic events related to subduction initiation and early arc development.

One of the most important characteristics linking SSZ ophiolites and forearc oceanic crust is the occurrence of boninites and arc volcanic suites overlying or accompa-nying MORB-like basaltic lavas (FIG. 2; Dilek and Furnes 2014 this issue). These spatial and temporal relationships imply that boninite magmatism, which involves fl uids derived from a sinking oceanic slab, occurred contempora-neously with or immediately after decompression melting that was associated with seafl oor spreading during the evolution of the IBM forearc crust and the SSZ ophiolites.

There are, however, some differences in detail between the forearc stratigraphy of the IBM arc and the internal structure and stratigraphy of SSZ ophiolites. For example, it appears that all volcanic rocks in the Troodos ophio-lite (Cyprus) show the infl uence of subduction. The lower pillow lava unit (LPL), which represents the fi rst volcanic sequence in the Troodos extrusive section, is composed of tholeiitic basalt and andesite with LILE enrichment caused by slab-derived fl uids (FIG. 5A; Bednarz and Schmincke 1994). This extrusive section shows increasing LILE/HFSE and LILE/REE ratios and decreasing HFSE and REE contents over time, implying an increasing contribution of slab-derived fl uids and depletion of the mantle source through time (Bednarz and Schmincke 1994). Thus, although the Troodos extrusive rocks are similar to the IBM forearc lavas at DSDP sites 458 and 459, they lack FABs with minimal slab infl uence, as are found at IBM dive sites.

The FAB-like basalt in the Semail (Oman) ophiolite (Geotimes Formation) is topped by low-Ti, depleted basaltic rocks that are free of slab-derived components (FIG. 2; Godard et al. 2006). These younger basaltic rocks are the products of a highly depleted mantle source, and their formation appears to have been followed by the eruption of lavas with subduction-affected compositions similar to those recovered from DSDP sites 458 and 459 (e.g. Reagan et al. 2010; Goodenough et al. 2014 this issue). Ishikawa et al. (2005) postulated that fl uids released by the dehydration of an amphibolitic crust, with a different composition from that of fl uids associated with a mature arc (for example, lower Ba/Rb, lower Pb), were added to the shallow mantle source where the boninitic magmas of the Semail ophio-lite were produced. We infer that this magmatism likely

FIGURE 4 Compilation of 40Ar/39Ar, K–Ar,

and U–Pb zircon ages for igneous rocks from the IBM forearc, modifi ed after Ishizuka et al. (2011a). The oldest products are 48–52 Ma forearc basalts, followed by ~44–48 Ma boninite and 43–45 Ma transitional high-Mg andesite. Arc tholeiites, repre-senting a mature magmatic arc, erupted at 44 Ma. The 40Ar/39Ar ages are calculated using 27.5 Ma for sanidine from the Fish Canyon Tuff (FC3). For data sources, see Ishizuka et al. (2011a, b).

49.3

44.0

DSDPSite 458Site 459

Bonin IslandsGuam Saipan

41.84

43.1–44.0(K-Ar)

40 Ma

45 Ma

50 Ma

South

42–35

35 Ma

Forearc basalts

Transitional suites(high-Mg andesite)

felsic rocks

Arc tholeiites and calc-alkaline lavas

35–36

45.1

?

41.0(K-Ar)

NorthBoninRidgetrench slope

51.63, 51.72(U–Pb: gabbro)

BoninRidgescarp

55 Ma

48.2–52(basalt)

?

?

(ODPsite 792)40.4, 40.6

42.3(ODPSite 793)

OmachiSeamount

31.79, 37.29

45.8

35.51, 37.2

51.1

Izuforearc

(ODPSite 786)Boninite and its differentiates

Ages are in millions of years (Ma). Data are 40Ar/39Ar ages unless otherwise noted.

46.0–48.2

44.3–44.7443.94–46.02

43.8842.86

Transitional forearc basalts

? 45.3–46.7

ELEMENTS APRIL 2014119

occurred in an immature, hot, shallow subduction-zone, analogous to the subduction initiation rock record along an oceanic plate boundary (Reagan et al. 2010).

Dilek and Thy (2009) reported the occurrence of sheeted dikes and extrusive rocks in the Kizildag ophiolite in southern Turkey with a geochemical signature charac-teristic of a depleted mantle source modifi ed by slab-derived fl uids. Boninitic magmas, produced by a high degree of melting of an ultradepleted mantle, appear to have formed contemporaneously with or shortly after the seafl oor spreading–related magmas of the sheeted dikes and extrusive rocks in this ophiolite (Dilek and Thy 2009); this relationship is reminiscent of the IBM forearc crust.

These observations from ophiolites formed along the Tethyan margin point to a magmatic evolutionary path that is similar to that of IBM forearc crust. Decompressional melting of a depleted MORB-type mantle was followed by melting of an even more depleted mantle to which slab-derived fl uid/melt was added. The temporal relationships among MORB-like, boninitic, and arc magmatism may vary in detail, but the formation of MORB-like FAB almost always predates the other types of early-arc magmatism in SSZ-type Tethyan ophiolites (Dilek and Furnes 2009, 2011).

The temporal relationships between boninites and other subduction-related rocks in SSZ ophiolites are variable, but in general, MORB-like basalts predate boninitic rocks, which predate and/or are coeval with arc tholeiites and calc-alkaline lavas (Dilek and Furnes 2009, 2011). Recent results from the IBM forearc illustrate a similar pattern. FAB erupted fi rst, followed by transitional and then boninitic lavas. Thereafter, transitional high-Mg andesites erupted simultaneously with relatively normal arc lavas in different

parts of the arc. This period was followed by the eruption of entirely subduction-related tholeiitic and calc-alkaline lavas. At a given location in the IBM forearc, the magmatic progression is generally from MORB-like basalt to boninite to arc tholeiites and calc-alkaline lavas. There is no evidence for a signifi cant time gap between the eruptions of any of these magma types in the IBM system. A similar progres-sion of magmatism in the Tethyan ophiolites and the IBM forearc implies that SSZ oceanic crust formation in both modern and ancient examples was associated with subduc-tion initiation in a forearc environment (Reagan et al. 2010; Dilek and Furnes 2011, 2014).

Similarities between the oceanic lithosphere of forearc settings and SSZ ophiolites also extend to the upper mantle units, which are composed of extremely depleted peridotites. For example, peridotites underlying gabbroic and basaltic rocks in the Papuan Ultramafi c Belt (PUB) in Papua New Guinea consist mainly of refractory harzbur-gites (Jaques and Chappell 1980). The extrusive units in the PUB include low-TiO2 basalts, which were produced from partial melting of this depleted mantle (Jaques and Chappell 1980). Depleted harzburgites have also been reported from many ophiolites (e.g. Dilek and Furnes 2009, 2011; Morishita et al. 2011b). The geochemically depleted characteristics of ophiolitic peridotites compared to upper mantle rocks from mid-ocean ridges might be caused by: (1) additional melt extraction during subduction initia-tion from an already depleted mantle; (2) a high degree of melting, aided by fl uid fl ux from a subducting slab; and (3) interaction between depleted basaltic/boninitic melt and the peridotites (Dilek and Furnes 2014). Ophiolitic perido-tites vary in the extent of their depletion, from extremely depleted harzburgite–dunite assemblages to moderately

MO

RB-

norm

aliz

ed

A BoninMariana

Forearc basaltBoninitePost-44 Maarc basalt

0.1

1

10

100

RbBa Th U NbTa K La CePbNd Sr SmHf Zr Ti EuGdTb Y Yb Lu

Ti (ppm)/1000

V (p

pm)

50

2010

0

100

200

300

400

500

0 5 10 15

PhilippineSeaMORB

Site 1149Pacific crust

C

MarianaFAB

Chichijimaboninite

PacificMORB

BoninFAB

TroodosLPL

TroodosUPL

Hahajima(<44 Ma)

B

0.1

1

10

100

La Ce Pr Nd PmSmEu Gd Tb Dy Ho Er Tm Yb Lu

TroodosLPL

TroodosUPL

BoninMariana

Forearc basaltBoninitePost-44 Maarc basalt

chon

drite

-nor

mal

ized

Troodos(LPL)

LREE HREE

A B

CFIGURE 5 (A) N-MORB (Sun and McDonough 1989)–normal-

ized incompatible trace element patterns for fresh glasses of forearc basalts from the IBM forearc (Reagan et al. 2010; Ishizuka et al. 2011a). Examples of a boninite, a post-44 Ma arc basalt, and a basaltic andesite from the lower pillow lavas (LPL) of the Troodos ophiolite (Bednarz and Schmincke 1994) are also shown (for other data sources, see Ishizuka et al. in press). (B) Variations in rare earth element patterns in Bonin forearc crust (following subduction initiation), normalized to the chondrite composition of Sun and McDonough (1989). The forearc basalt patterns contrast strongly with the U-shaped boninite patterns. Data for the Troodos ophiolite are from a compilation by Dilek and Thy (2009). The upper pillow lavas (UPL) show U-shaped patterns similar to those of the Chichijima Island boninite, but also show LREE-depleted patterns. Other data sources: Taylor et al. (1994); Reagan et al. (2008); Ishizuka et al. (2011a, in press). (C) V–Ti systematics for volcanic rocks from the IBM forearc. For data sources, see Ishizuka et al. (in press).

ELEMENTS APRIL 2014120

REFERENCESBednarz U, Schmincke H-U (1994)

Petrological and chemical evolution of the northeastern Troodos extrusive series, Cyprus. Journal of Petrology 35: 489-523

Bloomer SH and 6 coauthors (1995) Early arc volcanism and the ophiolite problem: A perspective from drilling in the Western Pacifi c. In: Taylor B, Natland J (eds) Active Margins and Marginal Basins of the Western Pacifi c. Geophysical Monograph Series 88, American Geophysical Union, pp 67-96

Cosca MA, Arculus RJ, Pearce JA, Mitchell JG (1998) 40Ar/ 39Ar and K–Ar geochro-nological age constraints for the inception and early evolution of the Izu–Bonin–Mariana arc system. Island Arc 7: 579-595

Dilek Y, Furnes H (2009) Structure and geochemistry of Tethyan ophiolites and their petrogenesis in subduction rollback systems. Lithos 113: 1-20

Dilek Y, Furnes H (2011) Ophiolite genesis and global tectonics: Geochemical and tectonic fi ngerprinting of ancient oceanic lithosphere. Geological Society of America Bulletin 123: 387-411

Dilek Y, Furnes H (2014) Ophiolites and their origins. Elements 10: 93-100

Dilek Y, Thy P (2009) Island arc tholeiite to boninitic melt evolution of the Cretaceous Kizildag (Turkey) ophiolite: Model for multi-stage early arc–forearc magmatism in Tethyan subduction factories. Lithos 113: 68-87

Dilek Y, Thy P, Hacker B, Grundvig S (1999) Structure and petrology of Tauride ophiolites and mafi c dike intru-sions (Turkey): Implications for the Neotethyan ocean. Geological Society of America Bulletin 111: 1192-1216

Godard M, Bosch D, Einaudi F (2006) A MORB source for low-Ti magmatism in the Semail ophiolite. Chemical Geology 234: 58-78

Goodenough KM, Thomas RJ, Styles MT, Schofi eld DI, MacLeod CJ (2014) Records of ocean growth and destruc-tion in the Oman–UAE ophiolite. Elements 10: 109-114

Hickey-Vargas R (1998) Origin of the Indian Ocean-type isotopic signature in basalts from Philippine Sea plate spreading centers: An assessment of local versus large-scale processes. Journal of Geophysical Research: Solid Earth 103: 20963-20979

Ishikawa T, Fujisawa S, Nagaishi K, Masuda T (2005) Trace element charac-teristics of the fl uid liberated from amphibolite-facies slab: Inference from the metamorphic sole beneath the Oman ophiolite and implication for boninite genesis. Earth and Planetary Science Letters 240: 355-377

Ishizuka O and 10 coauthors (2006) Early stages in the evolution of Izu–Bonin arc volcanism: New age, chemical, and isotopic constraints. Earth and Planetary Science Letters 250: 385-401

Ishizuka O, Yuasa M, Taylor RN, Sakamoto I (2009) Two contrasting magmatic types coexist after the cessa-tion of back-arc spreading. Chemical Geology 266: 274-296

Ishizuka O and 9 coauthors (2011a) The timescales of subduction initiation and subsequent evolution of an oceanic island arc. Earth and Planetary Science Letters 306: 229-240

Ishizuka O, Taylor RN, Yuasa M, Ohara Y (2011b) Making and breaking an island arc: A new perspective from the Oligocene Kyushu-Palau arc, Philippine Sea. Geochemistry, Geophysics, Geosystems 12: doi: 10.1029/2010GC003440

Ishizuka O, Umino S, Taylor RN, Kanayama K (2014) Evidence for hydro-thermal activity in the earliest stages of intraoceanic arc formation: implica-tion to ophiolite-hosted hydrothermal activity. Economic Geology, in press

Jaques AL, Chappell BW (1980) Petrology and trace element geochemistry of the Papuan Ultramafi c Belt. Contributions to Mineralogy and Petrology 75: 55-70

Meijer A and 6 coauthors (1983) Chronology of volcanic events in the eastern Philippine Sea. In: Hayes DE (ed) The Tectonic and Geologic Evolution of Southeast Asian Seas and Islands: Part 2. Geophysical Monograph Series 27, American Geophysical Union, pp 349-359

Morishita T and 6 coauthors (2011a) Diversity of melt conduits in the Izu–Bonin–Mariana forearc mantle: Implications for the earliest stage of arc magmatism. Geology 39: 411-414

Morishita T, Dilek Y, Shallo M, Tamura A, Arai S (2011b) Insight into the upper-most mantle section of a maturing arc: The Eastern Mirdita ophiolite, Albania. Lithos 124: 215-226

Mukasa SB, Ludden JN (1987) Uranium–lead isotopic ages of plagiogranites from the Troodos ophiolite, Cyprus, and their tectonic signifi cance. Geology 15: 825-828

Reagan MK, Hanan BB, Heizler MT, Hartman BS, Hickey-Vargas R (2008) Petrogenesis of volcanic rocks from Saipan and Rota, Mariana Islands, and implications for the evolution of nascent island arcs. Journal of Petrology 49: 441-464

Reagan MK and 15 coauthors (2010) Fore-arc basalts and subduction initia-tion in the Izu-Bonin-Mariana system. Geochemistry, Geophysics, Geosystems 11: doi:10.1029/2009GC002871

Sharp WD, Clague DA (2006) 50-Ma initiation of Hawaiian-Emperor bend records major change in Pacifi c plate motion. Science 313: 1281-1284

Stern RJ (2004) Subduction initiation: spontaneous and induced. Earth and Planetary Science Letters 226: 275-292

Stern RJ, Bloomer SH (1992) Subduction zone infancy: Examples from the Eocene Izu-Bonin-Mariana and Jurassic California arcs. Geological Society of America Bulletin 104: 1621-1636

Stern RJ, Reagan M, Ishizuka O, Ohara Y, Whattam S (2012) To understand subduction initiation, study forearc crust: To understand forearc crust, study ophiolites. Lithosphere doi: 10.1130/L183.1

Sun S-S, McDonough WF (1989) Chemical and isotopic systematics of oceanic basalts: implications for mantle compositions and processes. Geological Society of London Special Publication 42: 313-345

Taylor RN, Nesbitt RW, Vidal P, Harmon RS, Auvray B, Croudace IW (1994) Mineralogy, chemistry, and genesis of the Boninite Series Volcanics, Chichijima, Bonin Islands, Japan. Journal of Petrology 35: 577-617

Umino S and 6 coauthors (2008) Origin of the sheeted dike complex at super-fast spread East Pacifi c Rise revealed by deep ocean crust drilling at Ocean Drilling Program Hole 1256D. Geochemistry, Geophysics, Geosystems 9: doi:10.1029/2007GC001760

Warren CJ, Parrish RR, Waters DJ, Searle MP (2005) Dating the geologic history of Oman’s Semail ophiolite: insights from U-Pb geochronology. Contributions to Mineralogy and Petrology 150: 403-422

depleted clinopyroxene-bearing harzburgites, which are associated with boninitic and MORB-type magmatism, respectively (e.g. Dilek and Thy 2009; Morishita et al. 2011a, b).

In Tethyan ophiolites, the crystallization ages of plagio-granite and dolerite dikes in seafl oor spreading–related oceanic crust plot within a narrow time window during the early–late Cretaceous (for example, 95.3 Ma for the Semail ophiolite [Warren et al. 2005]; 90–92 Ma for the Troodos ophiolite [Mukasa and Ludden 1987]; 91.8–91.6 Ma for the Kizildag ophiolite [Dilek and Thy 2009]; and 90–92 Ma for the Tauride ophiolite [Dilek et al. 1999]). This age distri-bution, combined with their geochemical fi ngerprints, suggests that the development of the Tethyan SSZ ophiol-ites during the Cretaceous was nearly coeval along the more than 2000 km long belt and that it involved the simulta-neous initiation of a subduction zone within the southern Tethys Ocean.

ACKNOWLEDGMENTSWe acknowledge fi nancial support from the Japan Agency for Marine-Earth Science and Technology (JAMSTEC) and the Atmosphere and Ocean Research Institute, University of Tokyo, for the research cruises. We thank the Institute for Materials Research, Tohoku University, and the Japan Atomic Energy Agency for providing irradiation opportuni-ties for 40Ar/39Ar dating. Ishizuka thanks the Japan Society for the Promotion of Science for supporting a Japan–UK joint research program. This study was supported in part by the JSPS Grant-in-Aid for Scientifi c Research (C) 22540473 to Ishizuka. Reagan acknowledges the NSF-MARGINS support (EAR0840862). We thank Y. Dilek for providing us an opportunity to contribute to this issue of Elements and for his editorial comments, and H. Furnes and anonymous reviewers for thoughtful reviews.