Postmagmatic Tectonic Evolution of the Outer Izu‐Bonin Forearc Revealed by Sediment Basin Structure and Vein Microstructure Analysis: Implications for a 15 Ma Hiatus Between Pacific Plate Subduction Initiation and Forearc Extension

Abstract International Ocean Discovery Program Expedition 352 recovered sedimentary‐volcaniclastic successions and extensional structures (faults and extensional veins) that allow the reconstruction of the Izu‐Bonin forearc tectonic evolution using a combination of shipboard core data, seismic reflection images, and calcite vein microstructure analysis. The oldest recorded biostratigraphic ages within fault‐bounded sedimentary basins (Late Eocene to Early Oligocene) imply a ~15 Ma hiatus between the formation of the igneous basement (52 to 50 Ma) and the onset of sedimentation. At the upslope sites (U1439 and U1442) extension led to the formation of asymmetric basins reflecting regional stretch of ~16–19% at strain rates of ~1.58 × 10−16 to 4.62 × 10−16 s−1. Downslope Site U1440 (closer to the trench) is characterized by a symmetric graben bounded by conjugate normal faults reflecting regional stretch of ~55% at strain rates of 4.40 × 10−16 to 1.43 × 10−15 s−1. Mean differential stresses are in the range of ~70–90 MPa. We infer that upper plate extension was triggered by incipient Pacific Plate rollback ~15 Ma after subduction initiation. Extension was accommodated by normal faulting with syntectonic sedimentation during Late Eocene to Early Oligocene times. Backarc extension was assisted by magmatism with related Shikoku and Parece‐Vela Basin spreading at ~25 Ma, so that parts of the arc and rear arc, and the West Philippine backarc Basin were dismembered from the forearc. This was followed by slow‐rift to postrift sedimentation during the transition from forearc to arc rifting to spreading within the Shikoku‐Parece‐Vela Basin system.


Introduction
In 2014, the International Ocean Discovery Program (IODP) conducted three closely related drilling expeditions using the R/V JOIDES Resolution to explore the Izu-Bonin-Mariana (IBM) arc system (Figure 1), and the magmatic processes related to subduction initiation. Expedition 350 (Sites U1436 and U1437) was the first expedition designed to ascertain "the missing half" of the subduction factory in the IBM rear arc (Busby et al., 2017;Tamura et al., 2015). Expedition 351 (Site U1438) drilled west of the Kyushu-Palau remnant arc ridge with focus on IBM arc origins (Arculus, Ishizuka, Bogus, Gurnis, et al., 2015). Expedition 352 (Sites U1439 to U1442) drilled the igneous outer Bonin forearc, related to subduction initiation (Reagan et al., 2015). The IBM system is the type locality to examine the accretion of oceanic crust immediately after the initiation of subduction, arc evolution, and continental crust formation in a suprasubduction zone (SSZ) setting (Rudnick, 1995;Stern et al., 2003;Stern & Bloomer, 1992;Tatsumi & Stern, 2006). This predominantly submarine convergent plate boundary extends for 2,800 km from the Izu Peninsula to Guam. It resulted from~52 Ma of subduction of the Pacific Plate beneath the eastern margin of the Philippine Sea plate . IODP expedition results were published recently, notably concerning the magmatic evolution (Brandl et al., 2017;Yogodzinski et al., 2018;Hickey-Vargas et al., 2018;Shervais et al., 2019), and the age of IBM volcanic rocks (Barth et al., 2017;Ishizuka et al., 2018;Reagan et al., 2019). In addition, the drilled cores provided (hemi)pelagic sedimentary-volcaniclastic successions and tectonic structures that bear information on the tectonic evolution of the outer IBM forearc. The biostratigraphic record within the related sedimentary basins provides temporal constraints on the fault activity, and revealed a~15 Ma hiatus between the formation of the igneous IBM basement and the onset of sedimentation within these basins.
In this study we investigate the overall structure of fault-bounded sedimentary basins and the related tectonic structures at IODP Expedition 352 drill sites in order to reconstruct the structural and tectonic evolution of the IBM forearc, in particular the~15 Ma time gap between Pacific Plate subduction initiation and forearc extension with related sedimentary basin formation. The results from this study also have implications for the evolution of the IBM system at a lithospheric scale and are implemented into an overall tectonic model. The new results from structural and tectonic studies, together with its magmatic and sedimentary inventory, facilitate a comprehensive reconstruction of forearc architecture in a SSZ tectonic setting and are therefore relevant to the interpretation of SSZ ophiolites worldwide.  Tectonic structures mainly resulted from brittle deformation that took place within and in the vicinity of fault zones. Timing of fault activity is basically derived from the stratigraphy within the adjacent sedimentary basins, described in detail by Robertson et al. (2018). These fault zones display a wide variety of related deformation structures including discrete faults, cataclastic shear zones, extensional fractures, and veins in which minerals, mainly derived from hydrothermal fluids, precipitated (e.g., Kurz et al., 2015;Reagan et al., 2015). The faults and related veins were created after the formation of the forearc crust. Extension veins also help to elucidate the deformation history of their host rocks and the associated fluid effects (e.g., Bons et al., 2012;Hilgers & Urai, 2002;Ramsay & Huber, 1983). We therefore analyzed calcite vein microstructures from drill cores at the Expedition 352 Sites U1439, U1440, U1441, and U1442. Microstructures obtained from these veins reveal the related deformation conditions, in particular differential stresses, deformation temperatures, and constraints on strain rates. Similar structures were also studied within the magmatic basement of the Amami-Sankaku Basin (ASB) in the northwest Philippine Sea, drilled during Expedition 351 at Site U1438. The techniques used were electron backscatter diffraction (EBSD) combining with stress piezometry of mechanically formed e-twins of calcite.

Geological Background
The approximately north-south trending IBM arc is related to the west dipping subduction of the Pacific Plate under the eastern margin of the Philippine Sea Plate (PSP) (Figure 1). From north to south the subducting Pacific Plate steepens from about 30°beneath Japan to nearly vertical below the Mariana arc Holt et al., 2018, and references therein). It is assumed that the IBM subduction zone began as part of a hemispheric-scale tectonic structure (e.g., a transform fault) of ancient, dense lithosphere in the western Pacific Ocean (Bloomer et al., 1995;Hall et al., 2003;Stern, 2004).

Igneous Basement
The timing of large-scale lithospheric subduction is constrained by the age of igneous rocks of the IBM forearc that started in the Eocene, at~52 Ma (Bloomer et al., 1995;Cosca et al., 1998;Ishizuka et al., 2006;Ishizuka et al., 2018;Reagan et al., 2019). The initial spreading generated forearc basalt (FAB) lavas at~52 Ma; later volcanism produced low-Si, then high-Si boninites (~51-46 Ma: Ishizuka et al., 2006, Ishizuka, Tani, et al., 2011Reagan et al., 2013 to form the proto-Bonin Ridge. Magmatic activity appears to have migrated inboard, with the FAB erupting closest to the trench (Sites U1440 and U1441) and the boninites (Sites U1439 and U1442 and the Bonin Ridge) erupting farther from the trench. Arc andesites erupted in the western part of the Bonin Ridge ( Figure 1) after about 46 Ma (Ishizuka et al., 2006). Ishizuka, Taylor, et al. (2011) argue that the crustal stratigraphy of the IBM forearc is oceanic crust overlain by boninitic and later arc lavas, while Stern and Bloomer (1992) and Ishizuka et al. (2006) argue that a broad swath of forearc crust formed by seafloor spreading after subduction initiation. Relating to IODP Expedition 352 results, Reagan et al. (2017Reagan et al. ( , 2019 have developed a hybrid of these models, in which FAB and low-Si boninite magmas were generated during an initial period of seafloor spreading, whereas the high-Si boninites erupted subsequently at discrete volcanic centers (i.e., the nascent arc) at 51.3 Ma. The rear arc basalts of the ASB (Site 351-U1438) are younger than the Sites U1439 and U1442 boninite, which suggests that they formed by renewed volcanism west of the Bonin Ridge, after initial spreading ceased in the forearc .
The Philippine Sea Plate (PSP) (Figure 1) is characterized by a complex tectonic and magmatic evolution (Hall et al., 1995). It is surrounded by transform faults and subduction zones. Initial seafloor spreading began around the time of subduction initiation at~52 Ma to form the initial back arc basin in the PSP (i.e., the West Philippine Basin) (Deschamps & Lallemand, 2002;Ishizuka et al., 2018;Seton et al., 2012;Whittaker et al., 2007;Wu et al., 2016). The ASB, in which Site U1438 was drilled, is situated north of the West Philippine Basin and west of the Kyushu-Palau Ridge. The ASB sole comprises about 1.5 km of sediment overlying igneous oceanic crust. The ASB basement commonly encompasses basaltic sheet flows and dikes of high-Mg, low-Ti, tholeiitic basalts showing variable alteration and veining (Arculus, Ishizuka, Bogus, Gurnis, et al., 2015;Arculus, Ishizuka, Bogus, & the Expedition 351 Scientists, 2015). The geochemistry of the basement lavas Yogodzinski et al., 2018) indicates derivation from mantle source rocks that were more melt depleted than those of typical mid-ocean ridges, similar to the IBM FABs . The basaltic basement of the ASB (IODP Site U1438) is currently interpreted as part of the initial basement produced during subduction initiation (Arculus, Ishizuka, Bogus, & the Expedition 351 Scientists, 2015), although ages of the Amami-Sankaku basement are concurrent with high-Si boninite volcanism on the Bonin Ridge , allowing an alternative explanation that this was the first IBM back arc basin .

Postmagmatic Faulting and Sedimentation
The arc and forearc crust that formed after subduction initiation was affected by later fault-related deformation and chemical/hydrothermal alteration owing to tectonic deformation at the IBM forearc. Extensionrelated asymmetric sedimentary basins (e.g., half-grabens) are developed at Sites U1439 and U1442 on the upper trench slope (Christeson et al., 2016;Kurz et al., 2015;Robertson et al., 2018) (Figure 2). The basins are bounded by west dipping normal faults along their eastern margins, accompanied by syntectonic (hemi)pelagic and volcaniclastic sedimentation. According to shipboard data, the lowermost sedimentary units at Sites U1439 and U1442 were tilted eastward by~20°, and the tilted beds were covered by subhorizontal beds (Reagan et al., 2015). Drill cores revealed discrete shear structures with dominant reverse to oblique reverse slip along subhorizontal fault zones. These were either reactivated as, or transected by, normal faults, oblique faults with a normal slip component, and strike-slip faults (Reagan et al., 2015). At Sites U1440 and U1441, on the outer forearc, the sedimentary basins are bounded by normal and oblique-slip/ strike-slip faults. The sedimentary fill was not significantly affected by tectonic tilting. Additional extensional, mainly fault-related structures are steeply dipping to subvertical mineralized veins and extension fractures.
Biostratigraphic constraints from calcareous nannofossils reveal initial sedimentation at~35 Ma . Tephra layers higher in deep-sea successions indicate that explosive dacitic IBM volcanism started around 28.6 Ma . Since the IBM igneous basement formed at 52-50 Ma , a~15 Ma depositional hiatus must therefore exist prior to the oldest known sediment deposits.
The sedimentary-volcanogenic evolution of the forearc basins, together with their tectonic and paleogeographic implications for the IBM forearc, had recently been discussed by Robertson et al. (2018) and by Kutterolf et al. (2018). Site U1439 has by far the best overall recovery of the sedimentary succession, with approximately 180 m of mainly pelagic sediments and tephras. This site can therefore be taken as reference site for the general upslope structure and related postmagmatic tectonics. Based on biostratigraphic constraints, three major time slices of fine-grained background sedimentation were defined by Robertson et al. (2018). Time Slice 1 ranges from the early Oligocene (ca. 34.44-32.92 Ma) to early Miocene (ca. 23 Ma), with sedimentation of nannofossil chalk, marl, or limestone, variably mixed with volcaniclastic and tuffaceous sediment (i.e., calcareous siltstone or sandstone) as well as IBM tephras since 16 Ma. The boundary between Time Slices 1 and 2 coincides with the change in sediment dip, from subhorizontal (Time Slice 2) to continuously increasing dip angles below 127 m below sea floor (bsf) (Time Slice 1). This boundary also marks the demise of the early IBM-derived volcanism. Time Slice 2 ranges from the early Miocene to the mid-Pliocene (ca. 23-4 Ma), with clay/claystone and mud/mudstone (variably mixed with silt/siltstone and sand/sandstone). Time Slice 3 ranges from mid-Pliocene to Holocene (ca. 4-0 Ma) with sedimentation of nannofossil ooze and minor mud (variably mixed with tuffaceous sediment and tephras from IBM and mainland Japan).
The succession at Site U1442A has many similarities to that at Site U1439, although thinner, and the Miocene succession is not as complete; no early Miocene nannofossils were identified .
For Site U1440, the basal sediments are Oligocene (32.92 Ma), that is, coeval with the basal sediments at Sites U1439 and U1442. A major hiatus ranges from the earliest Oligocene to the Miocene-Pliocene boundary. Age profiles differ considerably at Site U1441 with late Miocene basal sediments directly overlying the basement. Neither Oligocene nor early Miocene sediments were recovered. Middle Miocene sediments are also missing .

Samples and Methods
During IODP Expedition 352 samples were taken subsequently to shipboard drill core description. The sampling mainly focused on deformational structures (veins and fault rocks) from several drill sites (U1439, U1440, U1441, and U1442) (Figures 1 and 2). The supporting site survey data for Expedition 352 are archived at the IODP Site Survey Data Bank (http://web.iodp.tamu.edu/UWQ/). Samples were primarily taken from core intervals, which showed both (1) obvious shear deformation along faults zones (cemented cataclasites and fault breccias) and (2) zones characterized by obvious fluid activity in terms of wall-rock alteration and precipitation within extensional veins, gashes, and voids along shear fractures. Additional samples from IODP Expedition 351 were obtained from the Kochi Core Center (Japan).
Representative samples were cut from the working half of the drill cores into slices, with a maximum size of 10 × 5 cm. Deformation structures were cut parallel to the drill-core axis (being defined as Z axis). Polished thin sections, with a size of~27 × 46 mm, were used for optical microscopy. About 200 thin sections were studied. The locations of samples described are listed in Table 1.

Seismic Images
Seismic images across Expedition 352 sites from Christeson et al. (2016), complimented with interpreted faults, unconformities, and sedimentary bedding are displayed in Figure 2. The data are prestack migrated images, converted to depth using velocities of 1,500 m/s for the water column, 1,700 m/s for sediments, and the appropriate crustal velocities from a coincident seismic refraction profile (Christeson et al., 2016). The sediment-basement interface is indicated by a change from continuous reflectors to lower-frequency, disrupted reflectors, as confirmed, where possible, by the depth of the drilled igneous basement-sediment contact. Within the sedimentary succession we obtain a change from shallow, subhorizontal reflectors to more chaotic, dipping reflectors which corresponds to the boundary between Time Slices 1 and 2 (~27 Ma) of Robertson et al. (2018). We interpret normal faults bounding the sedimentary basins, and throughout the igneous basement.
The evaluation of differential stresses from calcite twin densities generally follows the methods described by Brandstätter et al. (2017). The density of deformation twins (i.e., number of twins per mm) can be used to estimate differential stresses (e.g., Ferrill et al., 2004;Friedman & Heard, 1974;Rowe & Rutter, 1990;Rybacki et al., 2011). Thin sections of thirteen representative samples hosting twinned grains were analyzed for twin width and twin density using a Keyence VHX-6000 digital photomicroscope and associated data analysis software. The mean twin width was determined from the sum of twin widths for each twin set of a calcite grain. The twin density was derived by counting the number of twins per grain, normalized to a unit length of 1 mm. The mean twin density value of each grain was used for the piezometry calculations.
Other piezometers, following the equations after Rowe and Rutter (1990), were not considered as these are very sensitive to small changes in twin density and therefore may not be suitable for application to naturally deformed rocks (for details, see Brandstätter et al., 2017).

EBSD Analysis of Vein Calcite
Crystallographic orientations of calcite grains in highly polished, oriented X-Z thin sections (with Z parallel to the drill core axis) were measured using a scanning electron microscope which was equipped with an electron back-scatter diffraction (EBSD) system (HITACHI S-3400N Type II with HKL Channel5) at Shizuoka University. Twenty kilovolt accelerating voltage and low vacuum mode (30 Pa) were used. For microstructural observations and analyses, 30 μm thick thin sections were polished using 1 μm diamond paste and colloidal silica for >5 hr. Phase maps were obtained using step sizes of 5 to 7 μm. HKL Channel5 software was used to process map data by removing single pixels that differed by >10°and extrapolating nonindexed pixels with the average orientation of neighboring pixels.
EBSD data were processed using the OIM Analysis software. For the determination of potential slip systems, misorientation axes for misorientation angles from 2°to 5°, 5 to 10°, and 10°to 15°were displayed as contoured inverse pole figures (IPF) with reference to the trigonal calcite crystal system (hexagonal scalenohedral crystal class) using the MATLAB© toolbox MTEX (Bachmann et al., 2010). For two default orientations (grains, subgrains, crystals, and crystallographic axes), the misorientation is the rotation required to rotate one set of crystal axes orientation into coincidence with the other (based on a fixed reference frame) (Zhao & Adams, 1988). Misorientation axes generally are crystallographically controlled, as misorientations (e.g., subgrain boundaries) result from intracrystalline deformation (e.g., dislocation glide) by the activation 10.1029/2019GC008329 Geochemistry, Geophysics, Geosystems of distinct slip systems. IPFs display the orientation of crystallographic axes or normal to a crystallographic plane relative to a reference axis. In this study, the reference axis is the drill core axis. A preferred orientation of distinct crystallographic axes and/or planes therefore indicates which slip systems, defined by a crystallographic plane and a misorientation axis, were activated during intracrystalline deformation.

Raman Spectroscopy
Raman spectra of minerals were carried out in confocal mode using a Jobin Yvon LabRam HR800 microspectrometer equipped with an Olympus BX41 optical microscope and a Si-based charged-coupled device detector at the NAWI Graz Geocenter, Institute of Earth Sciences, University of Graz. The instrumentation uses a 100 mW Nd-YAG laser (532 nm emission), a grating of 1,800 grooves/mm, and a slit width of 100 μm. The spectral acquisition time was set to 10-20 s for all measurements between 100 and 1,200 cm −1 .

IBM Forearc and Sediment Basin Structure
Despite the known differences in the mineralogical and geochemical composition of the magmatic basement, the IBM forearc crust structure is continuous from the Bonin Ridge to the trench, with changes in thickness but only minor changes in seismic wave velocity (Christeson et al., 2016;Takahashi et al., 2009). Sites U1440 and 1441 are located within small normal and oblique-slip/strike-slip fault-controlled basins, respectively, on the lower forearc slope at water depths of 4,447 and 4,775 m (Reagan et al., 2015;Robertson et al., 2018). The faults extend toward the trench axis (Kurz et al., 2015;Reagan et al., 2015;Robertson et al., 2018). Sites U1439 and U1442 were drilled on the upper forearc slope at water depths of 3,128 and 3,162 m (Reagan et al., 2015;Robertson et al., 2018).

Upper Slope Sites
The upper slope sites are located in an area of NW-SE trending asymmetric half-graben structures that embody fault-controlled, >2 km wide basins ( Figure 2). Bathymetric ridges that are bounded by west dipping normal faults separate these half-grabens ( Figure 2). Normal faults extend from the seafloor to depths of at least a few hundred meters into the magmatic basement and are prevalent near all of the Expedition 352 drill sites (Christeson et al., 2016). The seismic reflection images and the shipboard data show that the sedimentary lamination from 0 to approximately 127 m bsf at Site U1439 is layered (sub) horizontally, whereas the layers between 127 and 153 m bsf have dip angles between 10°and 14° (Reagan et al., 2015). This corresponds to the apparent dip direction displayed by seismic prestack time migrated images ( Figure 2). From 153 m bsf down to the contact with the igneous basement, dip angles range from 15°to 20° (Reagan et al., 2015). The sedimentary layering at Site U1442 generally dips gently eastward down to 75 m bsf. From 75 m bsf to the basement contact dip angles continuously increase up to 35° (Reagan et al., 2015).
Displacement along the confining normal faults can be calculated from the dip angles of the faults and the sedimentary layering, taking account of the maximum sediment thickness in each basin above the hanging wall cutoff, as indicated in Figure 3. Assumptions are that the faults are planar at the scale of the cross section and that before rotating, layering was initially subhorizontal and that the faults have the same orientation and dip. Parameters are as follows: a fault dip angle φ of~40°near CDP 66400 in Figure 2, an average bedding dip angle θ of 10°from shipboard measurements of sedimentary bedding planes, and a maximum sediment thickness x of 200 m near common depth point (CDP) 66375. Using the trigonometry displayed in Figure 3a (d = x/sinφ), we estimate that the fault displacement d is approximately 311 m.
Extension (elongation) (e) and the stretch factor β can also be calculated from the fault dip angle φ and bedding dip angle θ, using the following equations:

10.1029/2019GC008329
Geochemistry, Geophysics, Geosystems Using a fault dip angle of 40°results in an extension e of approximately 0.19 for the basin adjacent to Site U1439, that is, a stretch factor β of approximately 1.19. Any possible strike-slip component was not considered, as the shipboard fault and slickenside data display major normal sense of shear (Reagan et al., 2015). The elongation for the basin around Site U1442 was determined similarly. A fault dip angle φ of~39°near CDP 66700 was measured in Figure 2, as well as a bedding dip angle θ of 8°; this results in an east-west directed extension e of~0.16, and a stretch factor β 1.16. The bulk extension for the upslope Sites U1439 and U1442 therefore is in the range of 0.175.
The faulting-related strain rates can be estimated from the overall displacement, the stretching factor, and the biostratigraphic constraints. As described above, the finite bulk stretching factor calculated for the basins at Sites U1439 and U1442 is about 1.175 (e = 0.175). Assuming continuous fault slip over the full biostratigraphically documented time range of approximately 35 Ma gives a minimum strain rate estimate in the range of 1.58 × 10 −16 s −1 . An assumption of bulk fault slip between 35 and 23 Ma (Time Slice 1 as defined by Robertson et al., 2018), that is, within 12 Ma from beginning syntectonic sedimentation at 35 to the 23 Ma unconformity, gives a strain rate estimate in the range of 4.62 × 10 −16 s −1 .

Lower Slope Sites
The sedimentary basins at Sites U1440 and U1441 have an approximately symmetric graben geometry. The lateral distance between these two basins is approximately 1.5 km (see Reagan et al., 2015). The fault pattern at Site U1440 is quite irregular, whereas the basin boundaries at Site U1441 are well defined by east and west dipping bounding faults. A fault dip angle φ 1 of~32.5°near CDP 5450 was measured in Figure 2 for the limiting fault at the western basin margin; at the eastern basin margin the fault dip angle φ 2 is~33°near CDP 5200. The sedimentary layering is generally subhorizontal; in the depocenter, particularly toward the eastern basin margin, the sedimentary bedding is inclined, indicating synsedimentary/postsedimentary faulting and tilting. The bulk basin width is~1,300 m ( Figure 2c). Dip-slip displacements can be calculated from the fault dip angles and the depth, z, of the base of the sedimentary succession above the igneous basement, as shown in Figure 3b; the latter (z) is in the range of~285 m bsf. Accordingly, the dip slip displacement is in the range of 528 m (d 1 ) and 520 m (d 2 ) for the west and east bounding fault, respectively. The related overall extension is therefore given by the ratio The calculation using tilted blocks (upper slope sites) yields stretching of the crust in the entire block-faulted area in terms of regional extension. The calculation in the symmetric graben (lower slope sites), however, yields the stretching of the graben itself and is therefore very local. The regional stretching basically depends on how far away the next graben is from the graben under consideration. The lateral distance from the eastern margin of Site U1441 basin to the western margin of Site U1440 basin is approximately 1.5 km. In order to make the stretching factor at upper slope sites and lower slope sites comparable, half of the distance from Site U1440 to the next graben on both sides (approximately 1.5 km in total), were added to l 0 used in the 10.1029/2019GC008329 Geochemistry, Geophysics, Geosystems calculation above (l 0 = 261 m; l 0 regional = l 0 + 1.5 km =~1,750 m). The regional extension for the lower slope sites is therefore estimated from Δl 1 , Δl 2 and l 0regional , being in the range of 538, 416, and~1,750 m. This results in a regional extension e regional of~0.545. Assuming continuous fault slip over the full biostratigraphically documented time range of~35 Ma gives a minimum strain rate estimate in the range of 4.395 × 10 −16 s −1 . An assumption of bulk fault slip between 35 and 23 Ma (Times Slice 1 as defined by Robertson et al., 2018) gives a strain rate estimate in the range of 1.43 × 10 −15 s −1 .

Faults and Fault-Related Structures
Several of the IBM sites revealed fault zones with various kinematic features, including discrete faults and cataclastic shear zones (Figure 4), together with extensional fractures and veins with minerals that were mainly precipitated from hydrothermal fluids ( Figure 5). Sites U1439 and U1441 are situated close to the related sediment basin axis, whereas Site U1442 is situated on a topographic, uplifted basement ridge. Drilling at these sites penetrated several minor fault zones within the igneous basement, most of these being related to the major, basin-bounding normal, oblique, and strike-slip faults. Site U1440 is located off the sedimentary basin axis, and penetrated the normal fault terminating this basin to the west, as well as related secondary extensional structures.
Fault structures that were observed in the Expedition 352 forearc sites are documented in detail by Reagan et al. (2015) and are summarized below. These data are supplemented by postcruise microstructural analyses. Fault rocks within Fault Zone 2 comprise cohesive fault breccia with millimeter to centimeter sized host rock fragments embedded within a fine-grained matrix (<0.2 mm grain size) (Figure 5f). The matrix amounts to less than 20%.
Fault Zone 3 contains slightly cohesive to incohesive fault breccia (centimeter-sized fragments) and cataclasites, with friable centimeter sized fragments of host rock material, partly surrounded by fine grained fault gouge and an irregular fracture network. Geochemistry, Geophysics, Geosystems  ( (Figures 5c and 5e).

Site U1441
The general structure at Site U1441 is highly irregular at the sediment-igneous basement boundary and along the basin margins and is characterized by multiple minor faults and fracture zones. Fault damage zones comprise steeply dipping to subvertical multiple sets of slickensides with predominant left-lateral to oblique-normal left-lateral shear. Mineralized veins are steeply dipping to subvertical.
The lowermost interval of Hole U1441A (~180 m bsf) revealed a semiductile low-angle shear zone (Figure 4d). This shear zone is characterized by an ultramylonitic fabric with distinct carbonate clasts and clasts of amorphic volcanic glass embedded within an ultrafine-grained, sheared calcium carbonate matrix, as revealed by Raman Spectroscopy. Clasts reach up to 500 μm in size. Shear bands appear as subparallel sets, indicating top-down sense of shear (Figure 4d). The shear bands are transected by subparallel sets of inclined shear fractures, indicating normal displacement, too. This shear zone marks the contact between clinopyroxene-phyric basalt above and aphyric to sparsely clinopyroxene-bearing basalt below (see the geochemical stratigraphy described by Reagan et al., 2015;Shervais et al., 2019). FAB pieces that were recovered below and above this shear zone do not indicate comparable deformation or alteration.

Veins and Wall Rock Alteration
Vein structures and alteration textures that were observed during IODP Expedition 352 are documented in detail by Reagan et al. (2015) and are summarized below. These data are supplemented by postcruise microstructural and piezometry analyses.

Macroscale Vein Structures
Mineralized veins ( Figure 5) were observed at several of the IBM sites and comprise tension fractures, hybrid (tension and shear) fractures, tension gashes related to releasing bends and extensional step-overs that are distributed along distinct shear fractures. The vein dip angles are usually >45°. In general, the individual drill sites are not particularly characterized by any distinct vein type. Veins occur as single subvertical features (Figure 5a), single bent and tapered tension gashes (Figure 5b), or as conjugate hybrid mineralized fractures (Figure 5c). Some of these veins, like at Site U1440, occur as intersecting networks (Figure 5c). The related vein dip angles (from~25°to 50°, and from~65°to 90°, respectively) form two clusters. Other vein types are splayed extensional veins with horsetail arrangement, associated with distinct single sets of shear fractures with top-down displacement (normal sense of shear) (Figure 5d), or as complex multiple vein networks, resulting from hydrofracturing ( Figure 5e). In addition, distinct domains with cataclasites and fault-related hydraulic breccias comprise disintegrated host rock fragments that are cemented by mainly sparitic carbonate (Figure 5f).
Calcite, various zeolites, and clay minerals are the main vein fillings. Vein thicknesses vary from <1 to 15 mm. Calcite veins are usually filled with crystalline blocky (Mg-) calcite. A few wider (>15 mm) veins have antitaxial, zoned calcite fiber mineralizations that indicate incremental steps of extension and precipitation.

Vein Calcite Microstructures and Twin Density Piezometry
Each of the investigated host rocks contained blocky calcite veins with twinned grains. Boninites, FABs and samples from ASB hosted nine, two, and three veins, respectively, that were suitable for twin density measurements. The grains predominantly display one twin set, but locally two sets and rarely three sets. Following the classification of Burkhard (1993) Figure 6). Locally, offsets along intracrystalline microfractures are observed.
According to Burkhard (1993), Type II twins point to deformation temperatures between~150 and 300°C. While tapered twins are distinctive of deformation and distinguish from growth twins, bent twins as well as offsets along microfractures point to advanced intracrystalline deformation. Moreover, undulatory extinction and the formation of subgrains within calcite grains with tapered and curved twins (Figure 7) indicate intracrystalline, plastic deformation mechanisms (dislocation glide and dislocation creep) (e.g., Twiss, 1977;Wheeler et al., 2001).
Most of the detailed studies were performed on samples from the boninite sites (U1439 and U1442), mainly due to a higher vein recovery. Differential stresses at the boninite sites range from 34 to 150 MPa with an average of 88 ± 9 MPa (Table 2). Veins from U1440 and U1441 (FAB) and from Site U1438 (ASB) show average differential stresses of 69 ± 7 (39 to 87 MPa) and 91 ± 33 MPa (34 to 181 MPa), respectively (Table 2). Basically, no correlation of differential stresses with depth was observed ( Figure 8).
The results of iterative strain rate calculations are shown in Table 3, indicating that strain rates (at given differential stresses) are highly sensitive to temperature. Depending on the assumed deformation temperatures, derived from calcite twin morphology and fault-and vein-related alteration assemblages (see below), the strain rates range from 10 −18 to 10 −16 at 100°C, from 10 −16 to 10 −14 at 150°C, and from 10 −14 to 10 −12 at 200°C.

Fault-and Vein-Related Wall Rock Alteration
Deformation temperatures can supplementarily be assessed by wall rock alteration mineral assemblages within the fault zones and adjacent to mineralized veins. In Hole U1439C wall rock alteration is indicated by secondary clay minerals, different zeolite types, and mutable occurrence of calcite. Additionally, the presence of talc within the lower sections of Hole U1442 indicates zeolite facies (metamorphic) conditions in the range of 100°to 150°C. The overall alteration degree is low to moderate at these sites (see Reagan et al., 2015, for details).
The extent of alteration at Site U1440 is generally low. Secondary minerals are clay minerals (montmorillonite, interlayered montmorillonite-illite, and illite), calcite, and small amounts of zeolite (largely laumontite and phillipsite) (Reagan et al., 2015). Macroscopic foliation fabrics defined by clusters of chlorite were observed between 145.00 and 146.00 m bsf, 281.00 and 291.00 m bsf, and 358.00 and 369.00 m bsf. The alteration mineral assemblages indicate upper zeolite facies to prehnite-pumpellyite facies metamorphic conditions. Alteration mineral assemblages at Site U1441 comprise zeolites (phillipsite, merlinoite, chabazite, and analcime), smectite group clays, and minor calcite. These assemblages indicate moderate to high alteration with a tendency to decline downhole (Reagan et al., 2015). In general, the alteration at the upslope Sites U1439 and U1442 appears to be lower, compared to the downslope sites. This difference also applies to the metamorphic conditions and the related (deformation) temperatures, allowing for higher strain rates at similar differential stresses at the downslope sites.

EBSD Analyses
The samples selected for the EBSD analysis are blocky calcite veins from Expedition 351 Site U1438 (ASB rear arc), and from Expedition 352 Sites U1439 (boninite) and U1441 (FAB). Several calcite veins contain twins and show undulatory extinction, bent twins and multiple subgrain boundaries. Twins are usually overprinted by or interact with subgrains (Figure 7).
The internal calcite microstructure within Sample ASB-3 (Site U1438) is characterized by the formation of elongate subgrains with an average size of~38.5 μm and an aspect ratio from 1.5 to 10 (Figure 9).

Geochemistry, Geophysics, Geosystems
Misorientation angles range from 1°to 9°; only a few subgrain boundaries show misorientation angles exceeding 5°. The misorientation profile shows higher misorientation angles in the rim areas with reference to the starting point than within the internal grain domains. Major jumps in misorientation angle are coupled with low-angle grain boundaries. For the determination of potential slip systems, misorientation axes for misorientation angles from 2°to 5°, 5 to 10°, and 10°to 15°were displayed as IPF with reference to the trigonal calcite crystal system (hexagonal scalenohedral crystal class) (Figures 9-11). Sample ASB-3 shows a maximum in orientation distribution function (ODF) density (max) of 3.1 and 2.6, respectively, parallel to the crystallographic c-axis [0001] at low (2°to 5°and 5°to 10°) misorientation angles. IPF for misorientation angles from 10°to 15°are characterized by a maximum relative to one of the f-planes (1102).
The calcite grains in the vein Sample BON-9 (Site U1439) have sizes in the same range as sample ASB-3. The subgrains are generally elongated, with aspect ratios between 1.5 and 5, and an average grain size of~30 μm ( Figure 10). The subgrains in this sample appear to be more evolved in comparison to Sample ASB-3, with misorientation angles of up to 90°and the formation of discrete high-angle boundaries. The misorientation profile shows several domains of internally constant misorientation with reference to the starting point; these domains are separated by high angle misorientation boundaries. For low angles (2°to 5°and 5-10°), the IPFs show a maximum in ODF density (max) of 1.8 and 1.9, respectively, parallel to the crystallographic c-axis [0001]. In addition, the 5°to 10°IPF is characterized by a maximum in the area of (11 0 8). IPF for misorientation angles from 10°to 15°are characterized by a maximum normal to one of the r-planes (1 014).

Geochemistry, Geophysics, Geosystems
The intracrystalline calcite microstructures within Sample FAB-3 (Site U1441) do not differ significantly from the samples described above (Figure 11). Subgrains have elongate shapes with aspect ratios between 1.5 and 3; the average subgrain size is~40 μm. An incremental increase in misorientation angle is apparent along the misorientation profile ( Figure 11). The internal domains are separated by subgrain boundaries with misorientation angles of 5°to 10°. For low angles (2°to 5°), the IPFs reveal a maximum in ODF density (max) of 1.3 in the area of (110 8), whereas the 5°to 10°IPFs are characterized by a maximum close to (1 014). IPF for misorientation angles from 10°to 15°are characterized by a maximum at (1 014) and (110 8).

Discussion
Seismic reflection data at the scale of the IBM forearc upper crust, mesoscale structures within drill cores, and fault and vein microstructures provide an insight into the tectonic evolution of the outer IBM forearc. The tectonic structures indicate that IBM forearc deformation essentially took place after formation of the igneous basement. Postmagmatic extension triggered the formation of asymmetric sediment basins, notably the half-grabens at Sites 352-U1439 and 352-U1442 on the upper trench slope, and symmetric graben structures at Sites 352-U1440 and 352-U1441 closer to the trench, with localized shear along multiple sets of faults. Faulting was accompanied by syntectonic pelagic, hemipelagic, and volcaniclastic sedimentation.

Deformation Conditions Derived From Microstructure Analysis
Calcite twin piezometry yielded mean differential stresses in the range of 52 to 140 MPa in general (Table 2), with mean differential stresses of~88 MPa at upslope sites,~69 MPa at downslope sites, and~90 MPa at Site U1438, respectively. Deformation microstructures indicate that subgrain formation and mechanical twinning have interacted, that is, subgrains overprint twins and vice versa. Subgrain formation in particular can be related to the effects of either temperature-reduced CRSS or increased  Geochemistry, Geophysics, Geosystems  Table 1.

10.1029/2019GC008329
Geochemistry, Geophysics, Geosystems Figure 10. Results of EBSD mapping for Sample BON-9. For legend and explanation see figure captions at Figure 9. For exact sample locations see Table 1.

10.1029/2019GC008329
Geochemistry, Geophysics, Geosystems Figure 11. Results of EBSD mapping for Sample FAB-3. For legend and explanation see figure captions at Figure 9. For exact sample locations see Table 1.
The temperature limits are constrained by the calcite twin morphology (Type I and Type II twinning), indicating deformation conditions between approximately 125 and 250°C. The upper temperature limits are also constrained by alteration-related mineral assemblages that indicate zeolite to prehnite-pumpellyite metamorphic conditions. The activation of high-temperature slip systems in calcite ( Figure 12) can therefore be neglected with reference to these upper temperature limits.
The related differential stresses were derived from calcite twin piezometry (Table 2). Although calcite piezometers typically reveal a wide range of results (also in standard deviation), these differential stresses in the range of 50 to 140 MPa were by all means sufficient to activate these low-temperature-high-stress slip systems. By assumption of generally constant differential stresses during bulk deformation at several sites, the inferred subgrain misorientations suggest the activation of low-temperature slip systems, in particular the f-as well as the easier r-slip system ( Figure 12).
As deformation temperatures appear to be higher at the downslope sites, compared to the upslope sites (~25-50°C) as indicated by alteration mineral assemblages), the related strain rates may also slightly increase toward the trench. The derived mean differential stresses in the range of 70 to 90 MPa and deformation temperatures of~125°C and~150°C at upslope and downslope sites, respectively, result in strain rates in the range of 4 × 10 −16 s −1 at 125°C, and 4.5-5.5 × 10 −16 s −1 at 150°C (see Table 3).
In detail, the microstructures also indicate a variation in the dominant deformation mechanisms from Site U1438 toward Site U1441. At Site U1438 (ASB), the dominant calcite deformation mechanisms are twinning and f-slip. At Site U1439 and Site U1441, the dominant calcite deformation mechanisms are twinning and fslip and, in addition, r-slip. This indicates locally changing deformation conditions from the rear arc toward the forearc, probably depending on differences in the tectonic setting, and essentially for Sites U 1439 to U 1442 on the distance to the trench. In particular, at Site U1441 extension was accommodated by a (semi) ductile shear zone (Figure 4d) along the western margin of a graben-shaped sediment basin. The formation of chlorite at Site U1440 (Figure 4c) also shows that the alteration was accompanied by fluid infiltration and related hydration reactions. Shearing combined with these reactions therefore potentially facilitated shear zone weakening, allowing for an increase of strain rates toward the trench. This increase can, conversely, enable the activation of additional (high stress) slip systems. Figure 12. Diagram, after De Bresser and Spiers (1997) summarizing the intracrystalline deformation mechanism of calcite. The gray box displays the range of obtained differential stresses; the corresponding range of deformation temperature was acquired from different twin parameters according the classifications after Ferrill et al. (2004) and Burkhard (1993), and also the host rock alteration mineral assemblages described by Reagan et al. (2015). Mechanical e-twinning occurs at low CRSS and at proportional low temperatures. With increasing differential stresses and/or strain rates, f and r slip and subgrain formation is enabled.

Faulting and Formation of IBM Forearc Basins
Extensional faults and veins reveal the postmagmatic deformation within the Izu-Bonin forearc upper crust subsequent to the formation of its magmatic basement. Differential stresses, derived from calcite microstructures, were sufficient to exceed common tensile strength of the oceanic crust. This was potentially aided by hydrothermal fluid pressures as indicated by, for example, hydraulic breccias (Figure 5f), allowing crustal failure during incipient extension. Time constraints for the tectonic deformation of the IBM forearc are provided by the magmatic ages from the igneous basement and the biostratigraphic record of the sedimentary cover. Zircon ages (Ishizuka, Tani, et al., 2011;Reagan et al., 2013;Reagan et al., 2019)  Sr isotope data from calcite veins described above (Micheuz et al., 2018), indicate that vein calcite precipitated from seawater during late Eocene to Oligocene times. Accordingly, the related vein microstructures indicate that deformation is clearly postmagmatic, within the time range of the oldest recorded biostratigraphic age of the sediment cover at Sites U1439, U1440, U1441, and U1442 (late Eocene to earliest Oligocene, that is,~35 Ma; Robertson et al., 2018). This implies a~15 Ma hiatus between the formation of the igneous basement, forearc extension and the onset of (hemi) pelagic sedimentation . The IBM forearc in general is sediment poor as it is widely remote from continental margin mass wasting sedimentation (e.g., Reagan et al., 2010;Robertson et al., 2018;Stern et al., 2003). Following Robertson et al. (2018), the hiatus was either controlled by topographic isolation, as indicated by epiclastic volcanic material at the basement-cover contact at Site U1439, or sediments were bypassing the outer forearc to accumulate in the IBM trench. Fault-controlled extensional basins, as observed at Expedition 352 drilling sites, therefore allowed the preservation of the sedimentary record with minimal reworking  from~35 Ma onward. The sedimentary and structural record at Sites U1439 and U1442 also suggests that displacement along confining normal faults was incremental, with a first episode of fault movement with syntectonic sedimentation during Time Slice 1 (as defined by Robertson et al., 2018) from circa 35 to 23 Ma.
The boundary between Time Slices 1 and 2 coincides with the change of layering from subhorizontal above to inclined below 127 m bsf, respectively, at Site U1439. At Site U1442 ship-board data (Reagan et al., 2015) show that bedding dip angles change from subhorizontal to up to 35°from 75 m bsf toward the basement contact. This angular discordance ( Figure 13) can be correlated with an increase of dip angles at Site U1439 below 153 m bsf at~27 Ma. The biostratigraphic data documented by Robertson et al. (2018) therefore provide a good constraint for this episode of faulting and related block tilting at circa 27-30 Ma (rift stage in Figure 13). This faulting increment immediately precedes the opening of the Shikoku-Parece-Vela basins at 25 Ma (IIshizuka, Taylor, et al., 2011). It also coincides with a period of PSP regional extension that preceded focusing of extensional strain in the West Philippine backarc Basin, when spreading decreased and became more east-west directed (Deschamps & Lallemand, 2002). For this interval the age-depth plots for Site U1439 as described by Robertson et al. (2018) are characterized by a very steep to subvertical gradient indicating high sediment accumulation rates. Early to mid-Eocene sedimentation at Expedition 351 Site U1438, as documented from biostratigraphic data (Arculus, Ishizuka, Bogus, & the Expedition 351 Scientists, 2015;Barth et al., 2017) and U-Pb detrital zircon ages, is not necessarily inconsistent with the evolution described above, as Site U1438 is located closer to the West Philippine Basin backarc spreading center, so that subsidence and sedimentation could have commenced earlier along the rear arc-backarc transition zone.
At Sites U1439 and U1442 the greater part of the fault displacement therefore took place from circa 35 to 23 Ma, accompanied by synrift sedimentation (Figure 13). Tephras originally deposited horizontally show the same tilt as the associated sediment beds within Time Slice 1. The oldest tephra age within Time Slice 1 at Site U1439 is 32.3 Ma, whereas the youngest tephra age in Time Slice 1 is an ash dated at~27 Ma . Accordingly, a major increment of displacement and block tilting occurred at or immediately after 27 Ma. This was followed by waning displacement, with slow-rift to postrift sediment deposition during the transition from rifting along the margins to spreading at the center of the Shikoku-Parece-Vela Basin system at 25 Ma (Ishizuka, Taylor, et al., 2011) (Figure 13). At that time spreading migrated from the West Philippine Basin, succeeded by the formation of the Shikoku and Parece-Vela Basins (Deschamps & Lallemand, 2002;Faccenna et al., 2018) (Figures 1 and 13). These basins were subsequently formed by seafloor spreading during early and middle Miocene time, in a backarc position relative to the IBM subduction zone system (Kobayashi et al., 1995;Kutterolf et al., 2014) (Figure 13).
Our forearc basin analysis revealed a declining stretching gradient from the downslope sites (in particular Site U1441) toward the upslope sites (U1439 and U1442). Extension e is in the range of~0.55 at Site U1441, with related strain rates of 3.31 × 10 −15 to 9.66 × 10 −15 s −1 , in contrast to e~0.175 with strain rates in the range of 1.58 × 10 −16 to 4.62 × 10 −16 s −1 at Sites U1439 and U1442. Normal faults that were located closer to the retreating trench ( Figure 13) therefore underwent more stretching compared to the upslope extensional faults, which can be explained by flexure of the downslope forearc toward the retreating trench. In addition, the shearing at Sites U1440 and U1441 was accompanied by alteration, which resulted from fluid infiltration and related hydration reactions, potentially facilitating rheological weakening and an increase of strain rates during extension.

A Lithospheric Scale Tectonic Model Working Hypothesis
According to Reagan et al. (2019), subduction initiation at~52 Ma was accompanied by rapid trench retreat and asthenospheric upwelling, followed by slab dehydration and melting of the depleted upper plate mantle. An early embryonic arc formed at~51 Ma after termination of rapid trench retreat. Vast tectonic extension may therefore not be expected during a period of plate coupling, beginning about the time that the protoarc formed (i.e.,~51 Ma) ( Figure 13). Discrete reverse faults, being later overprinted by younger extensional structures, are inferred to be related to that period of coupling between the subducting Pacific Plate and the IBM protoarc. Coupling presumably ended about the time of initial forearc extension and related basin formation~15 Ma later.
With an assumptive average Pacific Plate subduction rate of 50 mm/year (e.g., Fryer, 1996;Fryer et al., 1990;Stern et al., 2003;Faccenna et al., 2009Faccenna et al., , 2018Gong et al., 2018;Holt et al., 2018;Kong et al., 2018), and a subduction angle of~30°along the northern IBM subduction segment (as indicated by seismic tomography, e.g., Jaxybulatov et al., 2013;Gong et al., 2018;Kong et al., 2018;Holt et al., 2018), the Pacific lower plate is expected to reach the 660 km mantle discontinuity within~12-13 Ma after subduction initiation. Bending of the plate at this juncture can help to accommodate subduction rollback inception and decoupling along the plate boundary (Faccenna et al., 2009). Numerical models by Faccenna et al. (2009Faccenna et al. ( , 2018 and Čížková & Bina, 2013, Čížková & Bina, 2015 also show that subduction rollback commences within 10 to 20 Ma after subduction initiation. IBM forearc extension is therefore inferred to be related to Pacific slab rollback and the resulting trench retreat (Figure 13), presumably starting during Late Eocene times. In this case, the lack of sediment accumulation between~50 and~35 Ma would reflect the lack of accommodation space in the forearc, so that sediments bypassed into the trench.
Although the extensional strain rates in the range of 10 −16 to 10 −15 s −1 appear to be low, these rates just represent initial upper crust rifting of the IBM forearc. Toward the arc and the backarc (West Philippine spreading center; CBF rift of Ishizuka et al., 2018), rifting was assisted by magmatism probably related to the Oki Daito Plume   (Figure 13), so that effective stresses dropped and the lithosphere was weakened locally (e.g., Koptev et al., 2017). Magmatism-assisted rifting consequently resulted in a rheologically decoupled lithosphere, where initial brittle deformation was accumulated in the upper crust and subsequent lithospheric necking occurred due to weak lower crust and lithospheric mantle. The "boninite sites" (U1439 and U1442) are therefore situated between the more intensely stretched trench-near downslope sites and the weakened arc/rear-arc/backarc sites. Accordingly, these boninite sites form a less stretched, neutral domain between the retreating trench and the spreading arc and backarc. A major detachment potentially commenced along the upwelling mantle zone and the related plutons that were located beneath the IBM arc and the rear arc ( Figure 13) as inferred by Stern (2010). This resulted in an increase in extensional strain that culminated in the rifting of the Ogasawara Trough during the Eocene time (Ishizuka et al., 2006) (Figure 13). Accordingly, spreading and arc magmatism ceased in the Amami-Sankaku/West Philippine Basin and along the Kyushu-Palau Ridge, respectively, as it was confined by this detachment, resulting in shutting down the arc (Stern, 2010). Subsequent spreading in the Shikoku Basin is likely to have left some of the original arc behind the Ogasawara Trough. The dismembered crust to the west of the Shikoku Basin, now exposed along the Kyushu-Palau Ridge, therefore represents an arc-derived allochthon that is now located along the western margin of the Shikoku-Parece Vela Basin system. Continuous backarc spreading resulted in the formation of the Shikoku-Parece Vela Basin from 25 Ma onward (IIshizuka, Taylor, et al., 2011).
Most relevant here are SSZ ophiolites that preserve oceanic extensional structures and related sediments. The tuffaceous sedimentary cover of the Jurassic Coast Range ophiolite in central California (e.g., at Del Puerto Canyon) (Robertson, 1989;Hopson, & Mattinson, & Pessagno, 1981, Hopson, Mattinson, Pessagno, & Luyendyk, 2008 is comparable to the tuffaceous sedimentary cover of the Izu-Bonin forearc Robertson et al., 2018). Unfortunately, the Coast Range ophiolite, like many other circum-Pacific, Cordilleran-type ophiolites, is too small in outcrop and overprinted by later active-margin tectonics to easily recognize primary ocean-floor extensional structures. Also, throughgoing ocean-floor extensional structures are likely to have been reactivated during emplacement, so destroying any evidence of primary extensional tectonics. More promising are Tethyan ophiolites that are commonly preserved as huge intact sheets that may preserve ocean floor structures. Grabens and half-graben formed parallel to the seafloor spreading center in the Late Cretaceous Troodos ophiolite (Varga & Moores, 1985;Allerton & Vine, 1990). Although not tuffaceous, metalliferous and pelagic sediments formed within ocean-floor half-grabens of similar scale and structure to those of the Izu-Bonin forearc (Boyle & Robertson, 1984), in some parts several Ma after magmatic basement formation (e.g., Robertson, 1977). Similarly, in the Late Cretaceous Oman ophiolite, rifts orientated parallel (or obliquely) to the spreading center, notably the Alley structure (Smewing, 1980), hosted deep-sea deposits including basaltic talus, metalliferous and pelagic sediments (Fleet & Robertson, 1980). In Oman, seafloor extensional (and transverse) structures were reactivated during ophiolite emplacement, associated with protrusion of serpentinized ultramafic rocks and subophiolite accretionary melange (Robertson & Woodcock, 1982). However, in contrast to the Izu-Bonin forearc these extension-related structures were short-lived because subduction did not continue and create a related volcanic arc as associated tuffaceous sediments (e.g., Pearce & Robinson, 2010;Robertson et al., 2012).
Key features that SSZ ophiolites may share with the Izu-Bonin forearc therefore include (1) basal metalliferous-oxide sediments; (2) a hiatus between igneous basement formation and its deep-sea sedimentary cover; (3) synmagmatic to postmagmatic extension, resulting in the formation of normal faults and related extensional structures, and fault-bounded graben or half-graben ( Figure 13); and (4) a sedimentary cover that includes ophiolite-derived talus and tuffaceous strata in small extensional basins.

Conclusions
The general structure of the fault-bound sedimentary basins at IODP Expedition 352 drilling sites reflects the post Late Eocene tectonic evolution of the Philippine Sea Plate.
Extensional fault-and vein microstructures reveal the postmagmatic deformation conditions within the Izu-Bonin forearc upper crust subsequent to the formation of its magmatic basement. Vein calcite microstructures indicate intracrystalline deformation within a low-temperature-high stress regime (T <~250°C; σ > 50 MPa). Mean differential stresses in the range of 70 to 90 MPa, calculated from calcite microstructures, were sufficient to exceed common tensile strength of the oceanic crust. Extension resulted in the formation of symmetric and asymmetric, fault-bounded basins with initial syntectonic sedimentation. In the area of IODP Expedition 352 drilling sites, the IBM forearc was stretched by~16-19% at the "boninite sites" and up to 55% at "FAB sites," at strain rates in the range of 10 −16 to 10 −15 s −1 . These rates are also indicated by the vein and fault zone microstructures.
The published magmatic ages from the IBM forearc basement, isotope data, and the biostratigraphic record from the cover sediments revealed a~15 Ma hiatus between the rapidly forming near-trench seafloor after subduction initiation around 52 Ma and subsequent tectonic forearc extension. We explain this 15 Ma 10.1029/2019GC008329 Geochemistry, Geophysics, Geosystems time gap by ongoing rollback of the Pacific Plate that triggered upper plate extension. This was accommodated by normal faulting accompanied by extensional mineralized vein formation from late Eocene to early Oligocene times onward.
By introducing a superordinate general plate tectonic model, extension and arc spreading are inferred to have been tectonically controlled by Pacific Plate subduction rollback. Downslope increase of stretch is inferred to be related to upper plate flexure toward the retreating trench. Near the arc, rear arc, and the backarc, extension was potentially assisted by magmatism, resulting in advanced lithospheric extension around 27-26 Ma with subsequent spreading of the Shikoku and Parece Vela Basins (~25 Ma). This period coincides with a major unconformity within the sedimentary basin sequences. Parts of the IBM arc and rear arc, as well as the West Philippine Basin, were sheared off from the IBM forearc forming an allochthon that was delaminated from the IBM arc and rear arc. Along the IBM forearc this goes along with waning fault displacement and slow-rift to postrift sediment deposition, and the transition from rifting along the margins to spreading at the center of the Shikoku-Parece-Vela basin system.
The overall IBM forearc architecture and development as inferred from this study may serve as a reference setting, in particular with regard to postmagmatic rift-related structures and a sedimentary cover that includes tuffaceous strata and will help with the identification and interpretation of a SSZ origin of ophiolites worldwide.