Extensional tectonics in the Honshu fore-arc, Japan: integrated results of DSDP Legs 57, 87 and reprocessed multichannel seismic reflection profiles

Abstract

Listen

Translate article

Summary Leg 57 drilling on the Honshu fore-arc revealed late-Oligocene arc volcanism at Site 439 just 90 km from the axis of the present Japan Trench, on a basement palaeo-high. This, the Oyashio landmass, is defined by outboard- and inboard-onlapping reflectors in the thick (up to > 2 km) inner trench-slope section. The Neogene section at Site 439 documents subsidence of the Oyashio landmass from a late-Oligocene position at or near wavebase, and subduction erosion was invoked by Leg 57 scientists to explain the anomalously narrow late-Oligocene arc-trench gap. Hole 584, drilled during Leg 87 just upslope from the mid-slope terrace above the outer edge of the Oyashio landmass, produced equally unexpected results. Though it failed to reach basement, penetration of 940 m of slope strata down to middle Miocene revealed below 200 m subbottom a progressive down-hole increase in dips to as much as 70°. Palaeomagnetic data indicate that dips are trenchward. Spot-coring in holes 500 m upslope and 700 m downslope revealed a similar peculiar geometry. However, lithostratigraphic and biostratigraphic markers are subhorizontal or only gently inclined trenchward when correlated between the three holes. Landward-dipping listric normal faults, not imaged in the seismic records available at the time of drilling, are hence required within the slope section. Numbers of such faults, giving step-wise displacements of limited overall vertical displacement, would explain the sub-horizontal markers. This extensional geometry is substantiated by healed, predominantly normal, microfaults which occur throughout the tilted section. Detailed structural analysis reveals that tilting occurred progressively from middle Miocene through earliest Pliocene, was preceded by development of a pervasive vein structure, and was both preceded and accompanied by extensional microfaulting. Microstructural observations, and the apparent restriction of kindred structures in other active margin cores to non-accretionary forearcs, lead us to ascribe the vein structure to an initial response of semi-lithified slope sediments to extension, rather than to hydrofracturing alone. The multichannel record on which Site 584 drilling was based showed slope-parallel reflectors, and gave no clue to the structure actually encountered with the drill. However, a line through Site 439 some 70 km upslope reveals landward-dipping normal faults, albeit without the rotations required to explain the geometry at Site 584. Subsequent reprocessing of a seismic record crossing the equivalent position of Site 584 40 km S shows considerable small-scale faulting like that around Site 439. This, in concert with the core-scale structural data, suggests tectonic jostling with an extensional fragmentation ranging from microstructure to blocks 300–500 m across. Such tectonism increases in intensity towards the mid-slope terrace until coherent reflections can no longer be resolved. Basement is either a Cretaceous-Palaeogene accretionary complex akin to the Shimanto Belt of SW Japan (more likely), or old crystalline rocks akin to those exposed onland in the Kitikami massif (less likely). Whatever its nature, low-angle landward-dipping reflectors (LDRs) are a common feature of the basement rocks. The listric landward-dipping faults generated in the slope sediments during Miocene-early-Pliocene subduction erosion may accordingly be explained in two ways. If the LDRs in the basement are old thrusts, they could conceivably have been reactivated as normal faults which propagated up into the cover as it accumulated. Alternatively, the faults in the cover may sole on low-angle detachment faults within the slope section. Present data only allow a first look at what must be a highly complex three-dimensional extensional setting.