Since the DSDP era, five legs had been drilled in the area before ODP Leg 186. They were DSDP Legs 56, 57, and 87, which were drilled in the forearc, and ODP Legs 127 and 128, which were drilled in the backarc of the Japan Sea. Figure F4 shows the map of the area with drilled holes across northeast Japan.
One of the primary results of drilling in the forearc before Leg 186 was the determination of the subsidence history of the forearc over ~20 m.y. and the discovery of a subduction erosion regime (von Huene et al., 1982), which is now considered an important process of plate subduction. A summary of the forearc drilling results according to von Huene et al. (1982) is (1) there were periods of explosive volcanism separated by a long Paleogene interval without volcanism; (2) massive Neogene subsidence of the forearc area occurred simultaneously with arc volcanism; and (3) in the Pliocene, steepening of the landward slope of the trench changed the slope basins from areas of rapid sediment accumulation to areas of erosion. Another surprise was the unexpected discovery of andesitic volcanic rocks at Site 439, 90 km inward of the trench axis, which indicated that there is an offset of ~200 km between the present arc and where Oligocene volcanism occurred.
The advancement of coring technology allowed us to increase the deep core recovery rate from <50% to ~70%. Using advanced piston coring, less disturbed sediment cores were recovered at nearly full recovery. We summarize here various results of Leg 186 that we intend to explain in terms of the tectonic history of the island arc system.
Figure F5 is a plot of ash records from Sites 1150 and 1151 as well as Leg 56 and 57 sites. It is apparent that there was a major increase in volcanic deposits at Site 1150 at ~3 Ma and a decrease in the most recent ~0.5 m.y. At Site 1151, the increase starts at ~4 Ma. Farther north (40.6°N), in Hole 438A, volcanism increased from ~5 until ~2 Ma.
Sedimentation rates at Sites 1150 and 1151 are given and plotted in Figure F6 after corrections for compaction, which are insignificant but are nonetheless incorporated. The overall feature is a relative low between 16 and 10-8 Ma, an increase to peak rates at 6-4 Ma, then a decrease to 3-2 Ma. This change of trend occurred as the stress field changed from tensional to compressional or from arc subsidence to uplift. If we compare the trend with Figure F5, the overall pattern is similar. The time difference of the peaks may be partly due to age uncertainties, though some could be real.
Chemical analyses of pore waters show that chlorinity gradually decreases with depth from ~550 mM at the top of Site 1150 to 500 mM at ~200 meters below seafloor (mbsf). From ~550 mbsf, values abruptly decrease with depth to reach a minimum of 350 mM at ~700 mbsf. Chlorinity at Site 1151 also decreases with depth, with somewhat different characteristics, but the magnitude of change is similar to that at Site 1150 (Fig. F7). A similar trend is observed in the magnesium, potassium, and alkalinity profiles at both sites. Kopf et al. (this volume) summarized that the profound excursion in gas and water profiles can be best explained by deep-seated processes. We will refer to these results as supporting evidence of water flux, which influences the mechanical plate coupling.
Japan Sea drilling has provided ages of basement basalts that range from 24 to 17 Ma (Tamaki, Suyehiro, Allan, McWilliams, et al., 1992). Together with stratigraphic constraints of subsidence history (Ingle, 1992) and marine geomagnetic data, Tamaki et al. (1992) proposed the evolution scenario of the backarc area to be in extension from 32 to 10 Ma, during which time the Japan Sea opened and formed between 28 and 18 Ma. The initiation of widespread uplift and the compressional regime occurred between 10 and 7 Ma, which accelerated at ~5 Ma and experienced extremely high rates of uplift (~500 m/m.y.) between 2 and 0 Ma.
Jolivet and Tamaki (1992) related volcanic events and tectonic phases. In particular, they pointed out the starvation of the volcanic activity between 10 and 7 Ma coincided with the stress field change presented above. The variation of ash layers with time (14-4 Ma) was found to be similar to the data from the forearc, suggesting the eruption source to be the Japan arc.
We follow Sato and Amano (1991), who summarized the geological history of the northeast Japan arc since ~22 Ma mainly based on onland stratigraphy across the arc between 38° and 39°N. They proposed to recognize four temporal stages in the development of the arc. The first is the rifting stage (~22-15 Ma), characterized by formation of the half-grabens, extensional tectonics, volcanism, and high rates of sedimentation, which coincided with the opening of the Japan Sea. The second is the backarc basin opening stage (15-13 Ma), during which time the backarc side subsided rapidly. Then the transitional stage (13-2.4 Ma) followed with a thermal subsidence substage (13-8 Ma) and transition to compressional substage (8-2.4 Ma). The fourth is the shortening deformation stage (2.4 Ma to present), when the stress field became compressional (Fig. F8).
Uyeda and Kanamori (1979) classified the subduction zones into two end-member types, namely Chilean and Mariana types. The former is strongly coupled and characterized by truly large earthquakes, shallow-dipping subduction slab, and no active backarc spreading. Uyeda (1982) reviewed the case of the northeast Japan arc and suggested that the arc changed from Mariana- to Chilean-type mode based on the subsidence history of the inner trench slope, which showed a shift from subsidence to uplift in the Pliocene. He suggested the depth change, if it represents the shallowing of the depth of the trench, may indicate the mode change. He also refers to supportive lines of evidence to recent stress change from tensional (21-7 Ma) to compressional in the direction of subduction (Nakamura and Uyeda, 1980) and to vertical movement change from subsidence (17-10 Ma) to uplift in the Quarternary (Sugi et al., 1983).
A simple model of plate interaction was proposed to explain contrasting subduction stress systems between northeast Japan and southwest Japan (Wang and Suyehiro, 1999). The northeast Japan stress field is basically east-west compressional over a wide area, whereas that of southwest Japan is arc-parallel compression. Such a feature can be explained by the difference in the dip angle and the length of the stress transfer zone, keeping other parameters the same. This model is supported by observations of the intraplate earthquake mechanisms and the length of the zone of interplate thrust earthquakes, which are proxies of the stress field and the plate locked zone. A locked zone may be considered a shear stress transfer zone.
Jarrard (1980) compiled subduction parameters around the globe and showed that the distance between the trench and arc is highly correlated (R = 0.91) to the plate dip angle. He obtained
where,
These subduction zones include arc ages of 6-226 Ma; age is only weakly correlated with gap distance (R = 0.44). On the same Pacific plate today, the dip beneath the Marianas is steep, whereas beneath Tohoku it is shallow. Therefore, global observation indicates that the present arc position is more strongly dependent on the present dip angle than on age. More importantly, the depth to slab from the arc can be roughly obtained by multiplying gap and tan (Dip I), which results in ~103 ± 13 km (neglecting the Philippine plate because the present slab has not reached the arc). Indeed, it has been commonly observed that at present the depth of the slab beneath the volcanic front is generally ~100 km over a large range of slab dips. We assume that was the case in the Neogene also. The depth of magma segregation has been constant since 20 Ma (Tatsumi et al., 1994). This suggests that if we can reconstruct the arc history, the subduction dip angle history can also be estimated.
According to the geologic history described above, the location of the volcanic front (VF) changed with time (Ohguchi et al., 1989; Sato, 1994). Figure F9 shows the migration of the volcanic front from Ohguchi et al. (1989). The VF existed at the eastern margin of the present Japan Sea at 30 Ma and moved to the Kitakami area and off Fukushima between 25 and 22 Ma. Then the VF moved to its present position at ~12 Ma. Ohguchi et al. (1989) suggest steepening of the subduction angle from the migration of the volcanic front (Ohguchi et al., 1989). Pollitz (1986) suggested the shift to compressional state may be related to the absolute motion change of the Pacific plate between 5.2 and 2.5 Ma (Pollitz, 1986).
Figure F10 shows the change in plate geometry over time required by the position of the VF and the fixed depth to the subducting slab. The effect of subduction erosion causing the trench axis to move inward is small compared with the overall geometry change. We will now proceed to show that these inferred changes in plate dip angle can explain the change in the stress field inferred from geological records. In order to do so, we need to refer to temperature-controlled behavior of the crust and mantle and surface effects of mechanical coupling.