TECTONIC SETTING

Forearc regions of nonaccretionary convergent plate margins may expose the mantle of the overriding plate deep on the inner slope of the trench because in such forearcs there is no or very little sediment added to the outer edge of the overriding plate during subduction (Fig. F4). Although the concept of tectonic erosion is still controversial, it is clear from the morphology of the southern part of the IBM forearc that some degree of tectonic erosion is taking place along the inner slope of the trench (Fryer et al., 1999, 2000). Recent multichannel seismic (MCS) data suggest that subduction of the Ogasawara Plateau may have scraped off a portion of the underlying forearc mantle in the southern Bonin forearc region (Miura et al., 2004).

The structures of the Izu-Bonin and Mariana forearcs differ from one another in several fundamental characteristics:

1. The Izu-Bonin system is essentially straight (Fig. F3), whereas the Mariana forearc has a broad curvature (Fig. F2).
2. The Izu Bonin system has a 50-km-wide ridge that runs along the entire length of the system at the base of the inner trench slope, whereas the inner slope of the Mariana forearc varies in morphology along strike (e.g., Stern and Smoot, 1998; Stern et al., 2004).
3. Distribution of serpentinite seamounts on the Izu-Bonin forearc is confined to the 50-km-wide ridge at the base of the inner trench slope. They lie along the top of this ridge and occur as discrete shoals spaced at irregular intervals along the ridge (see Taylor, 1992). By contrast, the serpentinite seamounts of the Mariana forearc are distributed across a broader zone from ~15 to 90 km arcward of the trench axis and lie almost exclusively along fault traces on the outer half of the forearc (Fig. F2).
4. The Izu-Bonin forearc has undergone a history of extension that resulted in the formation of an along-strike graben structure that is progressively more pronounced toward the south (Taylor, 1992). By comparison, the outer half of the Mariana forearc displays mainly along-strike fault lineaments north of ~21°N and a complex of extensional faults south of 20°N with a predominantly northeast trend and a subordinate northwest trend (Stern and Smoot, 1998). This conjugate faulting is most likely caused by extension resulting from the increase in curvature with time of the Mariana forearc (Fryer et al., 1985; Wessel et al., 1994; Stern et al., 2004). The change in structure has been suggested to reflect the change in direction of convergence from nearly orthogonal in the southeastern corner of the system to nearly parallel between 21° and 25°N. Extensional faulting in the Mariana forearc and the resultant structures are the most likely causes of the wider distribution of serpentinite seamounts on the Mariana forearc than on the Izu-Bonin lower trench slope ridge.

Swath bathymetry mapping of the Mariana forearc has revealed large fault-controlled graben structures related to the conjugate extensional faulting mentioned above. Some have throws on fault scarps of up to 4 km. Such deformation can provide access for seawater into lower crust and upper mantle rocks and can facilitate serpentinization of peridotites in proximity to the faults. Exposures of serpentinite have also been recorded from the deep inner slope of several other nonaccretionary forearcs, for instance, the inner slopes of the Tonga and Scotia Trenches. Wherever the upper mantle is exposed in these forearcs, serpentinites can form simply by seawater interaction with peridoties. We would expect little compositional difference between serpentinite formed by this process in either the abyssal or suprasubduction-zone environment (e.g., Fryer, 2002).

The fundamental difference between the serpentinites of the IBM forearc region recovered during Legs 125 and 195 and those recovered from other oceanic sites is that in the peridotites from suprasubduction-zone serpentinite seamounts, the IBM forearc peridotites have reacted with slab-derived fluids (and/or melts) resulting in compositions that reflect subduction zone processes. The muds from the serpentinite seamounts drilled during these legs indicate a suprasubduction-zone provenance for the peridotitic protolith and a slab source for the fluids (Saboda et al., 1992; Mottl et al., 2003; Savov et al., this volume).