Description of paleomagnetic samples.
Lateral migration of the Oshima and Sorachi-Yezo Belts within south central Hokkaido was quantitatively evaluated by means of paleomagnetic analyses in order to identify allochthonous blocks on the northwestern Pacific margin. The remanence stability of the Late Jurassic to Early Cretaceous voluminous igneous succession of the Kumaneshiri and Sorachi Groups and the overlying forearc sediments of the Cretaceous Yezo Group was evaluated through rock magnetic experiments. Twelve of the sites yielded characteristic primary components residing in mixtures of titanomagnetite and hematite having various mixing ratios. After an appropriate correction of inclinations’ shallowing of the post-depositional detrital remanent magnetization (pDRM) based on anisotropic acquisition experiments of the isothermal remanent magnetization (IRM), we confirmed significantly shallow inclinations even for the flattening-corrected data set, implying northward transportation after the emplacement. Based on comparisons to expected paleomagnetic directions calculated from contemporaneous reference poles, we conclude that the allochthonous blocks, including south central Hokkaido, migrated northerly during the Early Cretaceous. Previous investigations of paleomagnetism and numerical modeling of burial processes of sedimentary basins indicate that some crustal blocks in Hokkaido and NE Japan experienced delayed transportation and eventually amalgamated with the mother continent by the end of the Paleogene.
- allochthonous block
- lateral migration
- central Hokkaido
The main part of the longitudinal mountainous range of Hokkaido is called the Sorachi-Yezo Belt (Figure 1). It is characterized by the occurrence of a Jurassic ophiolite (the Lower Sorachi Ophiolite) and a Cretaceous forearc basin sequence (the Nitarachi–Yezo Sequence). This ophiolite-forearc basin sequence is underlain tectonostratigraphically by Cretaceous accretionary complexes, whose eastern margin is defined as the Idonnappu Zone. Well-preserved accretion constituents have long been targets of geological surveys aiming to elucidate the evolutionary process of a longstanding arc-trench system (e.g., [1–3]). An implicit assumption in these studies was that the east-verging accretion had occurred somewhere along the northeastern Asian convergent margin and had a close genetic relationship with the autochthonous Cretaceous volcanic arc in the Sikhote Alin, Russia.
However, recent paleomagnetic studies have cast doubt on part of the Mesozoic arc constituents having originated from a remote area. Tamaki et al.  found that the Upper Yezo Group (Campanian: ), distributed in the Urakawa area, preserves stable detrital remanent magnetization (DRM) characterized by a significantly shallow inclination. Compared to the expected geomagnetic directions in northeastern Asia, untilted and flattening-corrected paleomagnetic directions require northward transportation of the Urakawa area by as much as 3400 km since the Cretaceous. The DRM of the Middle Yezo Group in the Oyubari area in central Hokkaido (Figure 1c; Cenomanian/Turonian: ) shows fairly deep inclinations, suggestive of an autochthonous origin , and the presumed Cretaceous Yezo forearc basin in East Asia (e.g., ) may be divided into some blocks with quite different tectonic histories.
To unravel this paradox, we executed paleomagnetic analyses on the Kumaneshiri, Sorachi, and Yezo Groups around the southern part of central Hokkaido (Figure 1c) in the course of the present study. Comparing the flattening-corrected inclination values obtained from the volcanic arc components and forearc sediments, we verified the self-consistency of the migration hypothesis of the ancient arc-trench system. The authors submit a quantitative constraint on tectonic models of the East Asian convergent margin.
2. Geological setting and sampling
Voluminous Late Jurassic to Early Cretaceous igneous and volcaniclastic rocks associated with tuffaceous sedimentary rocks are distributed in the Oshima and Sorachi-Yezo Belts of Hokkaido (Figure 1b) and are called the Kumaneshiri and Sorachi Groups, respectively. They both originated from subaqueous volcanism and show a strong resemblance in stratigraphic succession. Figure 2 presents composite columnar sections of the late Mesozoic strata for selected survey areas. Takashima et al.  confirmed gravels of oolitic limestone and trachyandesite suffering subaerial oxidation in the lower part of the Nunobe Formation and deemed that part of the igneous rocks of the Sorachi Group was derived from island arcs. The presence of the oolitic limestone also implies sedimentation under a considerably warm climate.
Because secondary magnetization caused by harsh alteration overprinted the magnetism, there are only a limited number of previous paleomagnetic works on the Sorachi Group. Hoshi and Takashima  measured remanent magnetization in dolerite and basalt (pillow lava) samples and obtained a shallow mean inclination implying northward transportation after emplacement. Large scatter in their data hindered precise discussion, but then Kitagawa et al.  analyzed limestone, volcaniclastic rock, and andesite with wider spatiotemporal coverage and reconfirmed significantly shallower inclinations than were expected for the coeval mother continent.
In this study, we obtained core samples to measure paleo- and rock magnetism at five sites from the Kumaneshiri Group (KU01–05), 17 sites from the Sorachi Group (FU01–03, PA01–04, PE01–07, SH03–05) and nine sites from the overlying Yezo Group (PO01–04, SH01–02, SO01–02, TO01) as shown in Figures 1 and 2. The lithofacies for all the sites are summarized in Table 1. Cores 25 mm in diameter were taken from each site using an engine drill, and the individual cores were oriented with a Brunton compass mounted on an aluminum orientation table. Along survey routes, sampling sites were selected to ensure that the structural attitudes needed for the tectonic tilt correction of paleomagnetic directions were clearly defined on outcrops. In the laboratory, cylindrical specimens 25 mm in diameter and 22 mm long were cut from each core sample.
|PO01||Silty tuff (weakly bioturbated)||Ponbetsu River (Nanashi Stream)|
|PO02||Medium sand-size tuff||Ponbetsu River (Nanashi Stream)|
|PO03||Medium – fine sand-size tuff||Ponbetsu River (Nanashi Stream)|
|PO04||Calcareous nodule||Ponbetsu River (Nanashi Stream)|
|SH01||Acidic tuff||Shirikishimanai River|
|SH02||Very fine sandstone||Shirikishimanai River|
|SO01||Tuffaceous mudstone||Soshubetsu River|
|SO02||Tuffaceous mudstone||Soshubetsu River|
|TO01||Acidic tuff||Hobetsu (Tomiuchi)|
|KU01||Pillow basalt||Kabato Range|
|KU02||Pillow basalt||Kabato Range|
|KU03||Basaltic volcaniclastic rock||Kabato Range|
|KU04||Basaltic volcaniclastic rock||Kabato Range|
|KU05||Basaltic volcaniclastic rock||Kabato Range|
|FU01||Dolerite dike||Furano (Nunobe)|
|FU02||Pillow basalt||Furano (Nunobe)|
|FU03||Pillow basalt||Furano (Nunobe)|
|PA01||Pillow basalt||Panketeshimanai River|
|PA02||Dolerite dike||Panketeshimanai River|
|PA03||Pillow basalt||Panketeshimanai River|
|PA04||Pillow basalt||Panketeshimanai River|
|PE01||Pillow basalt||Penketeshimanai River|
|PE02||Dolerite sill||Penketeshimanai River|
|PE03||Pillow basalt||Penketeshimanai River|
|PE04||Volcanic sandstone||Penketeshimanai River|
|PE05||Volcanic sandstone||Penketeshimanai River|
|PE07||Acidic tuff||Penketeshimanai River|
|SH03||Oolitic limestone (turbidite)||Shirikishimanai River|
|SH04||Volcanic fine sandstone containing mud patch||Shirikishimanai River|
|SH05||Volcanic mudstone||Shirikishimanai River|
3.1. Basic methods
Bulk initial magnetic susceptibility was measured for all the specimens using a Bartington susceptibility meter (MS2). We conducted progressive thermal demagnetization (PThD) tests on selected pilot specimens from each site, except on samples that were too fragile for repeated heating, for which progressive alternating-field demagnetization (PAFD) testing was adopted. Natural remanent magnetization (NRM) was measured using a cryogenic magnetometer (760-R SRM, 2-G Enterprises) in a magnetically shielded room at Kyoto University and spinner magnetometers (SMM-85, Natsuhara-Giken; SSM-1A, Schonstedt Instrument) at Osaka Prefecture University. The PThD test was performed up to 680°C in air using a noninductively wound electric furnace with an internal residual magnetic field of less than 10 nT. The PAFD test was carried out stepwise up to 100 mT with a three-axis tumbler contained in μ-metal shield envelopes.
3.2. Demagnetization tests
Figure 3 depicts typical results of the PThD testing showing in situ coordinates. Focusing on the Pinneshiri area in the Kabato Range, for five sites in the Kumaneshiri Group (KU01–05), we found the stable components to have a converging trend on origin of the vector-demagnetization diagrams across a broad distribution of unblocking temperatures (TUB) up to 600°C after the northerly component was demagnetized at around 300°C. In the eastern part of the sampling region (the Nokanan area), we successfully isolated similar high-TUB components of remanent magnetization for five sites in the Sorachi Group (PE04, PE07, and SH03–05) and two sites in the lower Yezo Group (SH01–02). The directions of the characteristic remanent magnetization (ChRM) were calculated using a three-dimensional least squares analysis technique after .
Five site-mean ChRM directions obtained from the Kumaneshiri Group show normal polarity, and the precision parameter (κ) improves after tilt correction (Figure 4). Seven sites in the Sorachi and Yezo Groups in the Nokanan area are clustered into antipodal ChRM directions (Figure 4), which after polarity inversion, pass a positive reversal test at a 95% confidence level (
3.3. Identification of ferromagnetic minerals
3.3.1. Spectrum of coercive force (HC)
In order to identify carriers of magnetic components in the samples, we undertook isothermal remanent magnetization (IRM) experiments. Stepwise IRM acquisition was performed according to an analytical technique developed by Kruiver et al. . Figure 5 shows the linear acquisition plot (LAP) and gradient of acquisition plot (GAP) of the IRM acquired in direct magnetic fields of up to around 3 T. As shown by these examples, the plots generated from a majority of the IRM data can be matched by single magnetic components with relatively low B1/2 values (the field at which half the IRM saturation is reached), indicating the existence of low HC ferromagnetic minerals. On the basis of the TUB spectra mentioned before, we believe the remanent magnetization of the major samples resides in titanomagnetite.
3.3.2. Thermal demagnetization of orthogonal IRMs
We performed PThD of composite IRMs on selected specimens. Based upon the procedure proposed by Lowrie , composite IRMs were imparted by applying direct magnetic fields of 3.0, 0.4 and then 0.12 T to the specimens in three orthogonal directions. As shown in Figure 6, the decay curve of the IRM components derived from PThD testing indicates that the dominant magnetic phase is generally the low HC (<0.12 T) soft fraction with a broad spectrum of TUB up to 580°C. In such a case, the major carrier of the high-TUB component of the NRM is titanomagnetite. Smaller amounts of medium (0.12 < HC < 0.4 T) and hard (0.4 < HC < 3.0 T) fractions were identified. As for the basaltic rocks of the Kumaneshiri Group (KU0141, KU0231, and KU0511), a minor medium fraction is interpreted to be carried by fine (SD-size) grains of magnetite because they have TUB spectra up to 580°C. A small amount of the hard fraction in volcanic samples (PE0421 in the Sorachi Group, SH0111 in the Yezo Group) is carried by hematite because they have TUB spectra up to 680°C. These experiments clarified that the single-component NRMs preserved in samples of the Kumaneshiri, Sorachi and Yezo Groups are carried by a mixture of titanomagnetite and hematite mixed in various ratios.
3.4. Magnetic fabric
3.4.1. Anisotropy of magnetic susceptibility
In order to describe the magnetic fabric of the samples, we determined each specimen’s anisotropy of magnetic susceptibility (AMS). Measurement was done using a KappaBridge KLY-3S magnetic susceptibility meter (AGICO). All the results of AMS measurements are presented graphically in Figure 7 except for site KU01, for which there were too little data to obtain statistical parameters. The tilt-corrected AMS fabric (principal susceptibility axes) presented on an equal-area projection for the volcanic and volcaniclastic rocks (e.g., FU01–03, KU02–05, PA01–04, PE01–03, SH01) seems to be irrelevant to the sedimentary surface (horizontal plane). On the other hand, a majority of the sedimentary rocks (e.g., PE04–06, PO01–04, SH02, SO01–02) exhibit arrangements of AMS axes bound to the bedding plane, namely, the minimum axis (
3.4.2. Anisotropy in IRM acquisition
We performed inclination shallowing testing based on the method of  using IRM anisotropy. On one specimen per site, from which we obtained the ChRM, we applied a direct magnetic field at 45° to the bedding plane to avoid any field impressed anisotropy . We then measured the IRM, which was parallel (IRMx) and perpendicular (IRMz) to the bedding. Figure 8 shows typical IRM acquisition curves for IRMz and IRMx. The value of IRMz is lower than IRMx for the entire range of acquisition (Figure 8a), suggesting that ferromagnetic minerals carrying NRM are anisotropic and follow a similar trend. The ratio IRMz/IRMx can be uniquely related to the amount of inclination shallowing (tan I/tan IF = IRMz/IRMx; IF = inclination of the field in which remanence was acquired). The average IRMz/IRMx ratio for each sample was determined from the best-fit slope of IRMz against IRMx (Figure 8b).
4.1. Implications of shallowing correction
Inclination shallowing was found for most of the analyzed samples. As the shallowing occurred during postdepositional compaction prior to tectonic tilting, we applied the shallowing correction for the untilted data set in the study area. Figure 9 presents the primary magnetic directions of the Kumaneshiri, Sorachi, and Yezo Groups before and after shallowing correction. As suggested by Jackson et al. , anisotropy of IRM or anhysteretic remanent magnetization (ARM) may decrease as a result of postdepositional processes such as electrostatic and coagulation effects, whereas the direction of the primarily acquired pDRM is immune from such secondary effects. In that case, our flattening estimation, based on IRM anisotropy, may underestimate the actual amount of inclination flattening. Our corrected data set (Table 2) would then provide a minimum estimate of the paleolatitude. Based on the assumption that the data sets are records of the earth’s dipole magnetic field, the inclinations of formation-means for the Pinneshiri and Nokanan areas correspond to paleolatitudes of 20°N and 9°N, respectively.
|Site||N||D (°)||I (°)||DC (°)||IC (°)||IF (°)||α95 (°)||κ|
4.2. Possible allochthonous blocks in central Hokkaido
The present study ratified the hypothetical transportation from low latitudes proposed by the authors in Refs. [10, 11]. The central part of Hokkaido, however, does not seem to have migrated en bloc because the Mesozoic paleomagnetic records in some areas are indicative of autochthonous origin. Our PThD examinations for the Yezo Group distributed along the Ponbetsu River (PO01–04) imply deep inclinations, although they were not magnetically stable enough to determine site-mean ChRM directions. As with the Oyubari area studied by Tamaki and Itoh  (see Figure 1), we posit that the western wing of the N-S trending anticline may be composed of in situ blocks, whereas we obtained significantly shallow inclinations from the core and eastern wing of the structure. An unsolved problem is that a clear geologic boundary between the blocks with mixed origins has never been detected.
Dismemberment of amalgamated blocks during the late Cenozoic collision between the Kurile and northeastern Japan arcs is also a knotty problem for reliable paleoreconstruction. Exploration drilling was executed in 1997 in central Hokkaido (42.9943°N, 142.0228°E; ). The vertically drilled borehole reached 4465 m depth and confirmed that the Cretaceous to Tertiary strata was repeatedly stacked by remarkable west-vergent thrusts. Actually, the western foothills of the backbone mountains consist of multiple-stacked thrust horses and the flat-lying horst-graben of the Paleogene and Cretaceous igneous basement . Based on an interpretation of regional seismic profiles, Kazuka et al.  estimated the east-west crustal shortening across the Hidaka Mountains to be ~60 km. Thus, the origin of the allochthonous terranes should be clarified through further investigation of the three-dimensional structure of the island.
4.3. Estimate of N-S transportation
In order to determine the amount of tectonic movement of geologic units now distributed in south central Hokkaido, the expected direction should be calculated from the contemporaneous reference pole of the North China Block (NCB) after . Because it formed a single entity with Siberia and Mongolia around the Early Cretaceous , the whole block can represent a coherent part of East Asia since the Cretaceous time. Figure 10 presents a summary of Mesozoic to Cenozoic paleomagnetic information around the study area. Our results basically agree with the estimate of N-S transportation by the authors in Refs. [10, 11]. Based on temporal decreases in positive F values defined by Beck , allochthonous blocks migrated northerly during the Early Cretaceous. Paleomagnetic data reported from the Oyubari area in central Hokkaido  show an affinity to the NCB data. Considering a reconstructed subduction history by Ueda and Miyashita , simultaneous events of amalgamation on the continental margin may have occurred from 100 to 90 Ma.
Kimura et al.  regarded the Sorachi Group as a constituent of an enormous oceanic plateau that was driven to collide against the continent by rapid northward movement of the Izanagi Plate . It seems, however, that their hypothesis clashes with the evidence of arc volcanism, as mentioned above. Takashima et al.  proposed an alternative idea that the geologic unit was transported by left-lateral slips on the margin between the Eurasian and Izanagi Plates. Although oblique subduction is a plausible cause of the large migration, lateral motion on the plate margin changed to a dextral direction after the demise of the Izanagi Plate . In order to reconcile such controversial points, we present a comprehensive paleoreconstruction in the last chapter of this book.
Paleomagnetic analysis made it clear that some crustal blocks experienced delayed transportation. A composite mean of the Upper Yezo Group in the Urakawa area  gives a paleolatitude of 16.7°N. Using the NCB expected direction for comparison, northward transportation since the Late Cretaceous appears to be 3400 km. Based on geochemical modeling, Itoh et al.  suggested continued subsidence of the allochthonous ‘forearc’ region containing Urakawa area through the Paleogene. The considerable thickness of the missing unit is not attributed to eustatic sea-level changes but to tectonic subsidence of the forearc, which implies that a subduction erosion process  was active. It is noteworthy that an autochthonous block in central Hokkaido also suffered Paleogene subsidence. Tamaki et al.  executed 1D basin modeling on the basis of organic maturation data obtained from a deep borehole (MITI Yubari; ). Their burial history was better constrained because the data set contained maturity levels and the present thickness of the Eocene sedimentary units. They found accelerated accumulation rates during the Paleogene, which is an indicator of the emergence of a foreland basin setting (e.g., ). Simultaneous inversion of these areas implies that an amalgamation of migrated terranes occurred by the end of the Paleogene.
Lateral migration of the Oshima and Sorachi-Yezo Belts within south central Hokkaido was quantitatively evaluated by means of paleomagnetic analyses. We tested the remanence stability of the Late Jurassic to Early Cretaceous voluminous igneous succession of the Kumaneshiri and Sorachi Groups and the overlying forearc sediments of the Cretaceous Yezo Group through rock magnetic experiments. Twelve of the sites yielded characteristic primary components residing in mixtures of titanomagnetite and hematite having various mixing ratios. Although anisotropic acquisition of the isothermal remanent magnetization suggested shallowing of inclinations of the post-depositional detrital remanent magnetization, we confirmed significantly shallow inclinations even for the flattening-corrected data, implying northward transportation after emplacement. Based on comparisons to expected paleomagnetic directions calculated from contemporaneous reference poles, we concluded that the allochthonous blocks, including central Hokkaido, migrated northerly during the Early Cretaceous. Previous research studies of paleomagnetism and numerical basin modeling of burial processes indicate that some crustal blocks experienced delayed transportation and eventually amalgamated with the mother continent by the end of the Paleogene.
The authors extend their appreciation to Naoto Ishikawa for use of the rock magnetic laboratory at Kyoto University. We thank Yuki Bito and Takashi Inoue for their earnest help during field work and experiments.