Tectonic evolution of the triple junction off central Honshu for the past 1 million years

Abstract

We discuss several models of the evolution of the trench-trench-trench triple junction off central Honshu during the past 1 m.y. on the basis of plate kinematics, morphology, gravity and seismic reflection profile data available for the area. The study area is characterized by large basins, 7–8 km deep on the inner lower trench slope on the Philippine Sea side and the deep (9 km) Izu-Bonin Trench to the east. Between the basins and the trench, there are 6–7 km-deep basement highs. The triple junction is unstable due to the movement of the Philippine Sea plate at a velocity of 3 cm/yr in WNW direction with respect to Eurasia (Northeast Japan), subparallel to the strike of the Sagami Trough. Generally we can expect the boundary area between the Philippine Sea and Pacific plates to be extended because the Pacific plate is unlikely to follow the retreating Philippine Sea plate due to the obstruction of the southeastern corner of Eurasia. The above peculiar morphology of the junction area could have resulted from this lack of stability. However, there are several possible ways to explain the above morphology. Our gravity model across the trench-basement high-basin area shows that the basement highs are made of low-density materials (1.8–2 g/cm 3). Thus we reject the mantle diapir model which proposes that the basement highs have been formed by diapiric injection of serpentinites between the retreating Philippine Sea plate and the Pacific plate. The stretched basin model proposes that the basins have been formed by stretching of the Philippine Sea plate wedge. We estimated the extension to be about 10 km at the largest basin. We reconstructed the morphology at 1 Ma by moving the Philippine Sea plate 20 km farther to the east after closing the basins, and thus obtained 8 km depth of the 1 Ma trench, which is similar to that of the present Japan Trench to the north. Although this stretched basin model can explain the formation of the basins and the deep trench, other models are equally possible. For instance, the eduction model explains the origin of the basin by the eduction of the Philippine Sea basement from beneath the basement high, while the accretion model explains the basement highs by the accretion of the Izu-Bonin trench wedge sediments. In both of these models we can reconstruct the 1 Ma trench depth as about 8 km, similar to that of the stretched basin model. The deformation of the basement of the basins constitutes the best criterion to differentiate between these models. The multi-channel seismic reflection profiles show that the basement of the largest basin is cut by normal faults, in particular at its eastern edge. This suggests that the stretched basin model is most likely. However, the upper part of the sediments shows that the basement high to the east has been recently uplifted. This uplift is probably due to the recent (0.5 Ma) start of accretion of the trench wedge sediments beneath this basement high.