Seismic structures of the 154–160 Ma oceanic crust and uppermost mantle in the Northwest Pacific Basin
- First Online:
- Cite this article as:
- Oikawa, M., Kaneda, K. & Nishizawa, A. Earth Planet Sp (2010) 62: e13. doi:10.5047/eps.2010.02.011
- 250 Downloads
We present detailed P-wave velocity models for the Northwest Pacific Basin which were produced in 154–160 Ma at a high seafloor spreading half-rate of >8 cm/yr and have not been appreciably deformed by tectonic or igneous activity since then. We carried out wide-angle seismic experiments on two crossing survey lines which are respectively parallel and perpendicular to paleomagnetic lineations. The seismic crustal models for both lines are almost identical and homogeneous along these lines. The crust consists of an upper layer (Layer 2) with a P-wave velocity Vp = 2.5–6.8 km/s and a thickness of 1.3–2.2 km, and a lower layer (Layer 3) with a velocity of 6.8–7.1 km/s and a thickness of 4.6–5.9 km. These characteristics indicate that the crust beneath the survey line has a standard oceanic crustal structure. The structure of the uppermost mantle of the line parallel to the seafloor spreading direction exhibits considerable Vp heterogeneity within 5 km immediately below the Moho and shows an unusually high Vp of 8.5–8.7 km/s. The Pn velocity for the perpendicular line is 7.9 km/s, and the magnitude of the velocity anisotropy for the uppermost mantle amounts to a large value of 7–10%.
Key wordsOceanic crust Northwest Pacific Basin seismic structure ocean bottom seismograph anisotropy
Oceanic crusts created at mid-ocean ridges are considered to have similar basic structure everywhere, with the exception of anomalous regions such as fracture zones and hotspots (White et al., 1992). Most of the structural models of normal oceanic crust were obtained through seismic studies conducted between 1960 and 1980 (e.g., Shor et al., 1970; Christensen and Salisbury, 1975). Recently, the remarkable advancement of technologies for studying seismic refraction and reflection (e.g., GPS navigation systems, Ocean Bottom Seismographs (OBS), analytical methods, calculation performance of computers) has allowed for the acquisition of more precise velocity structures (e.g., Kasahara et al., 2007; Oshida et al., 2008). However, there have been only few seismic explorations aimed at studying normal oceanic crust with the latest technologies.
Accordingly, we carried out two wide-angle seismic experiments on the two survey lines in the area, which were designed to obtain accurate and updated models of the typical oceanic crust and uppermost mantle (Fig. 1).
2. Seismic Experiments and Data Processing
The seismic experiments presented here were conducted using the S/V SHOYO and S/V TAKUYO vessels in 2006 and 2007 under the Continental Shelf Surveys Project implemented by the Hydrographic and Oceanographic Department, Japan Coast Guard. We focused on two survey lines, MTr6 and MTr8 northwest of Minami-Tori Shima Island (Fig. 1). MTr6 and MTr8 run perpendicular and parallel to the paleomagnetic lineation, respectively. MTr6 intersects MTr8 perpendicularly 40 km from the northwestern end of MTr6. Since many small seamounts reflecting significant past igneous events are scattered in the eastern part of MTr6, we do not discuss this area in this paper.
For a controlled seismic source for wide-angle seismic experiments, we used an array of four air guns with a total volume of 96 ℓ (6,000 in3) firing at intervals of 200 m. One hundred OBSs were installed at every 6 km along MTr6 and 17 OBSs were installed at every 10 km along MTr8. The positions of the OBSs were determined with the global search method (Oshida et al., 2008). Single-channel seismic reflection (SCS) data were also collected along the survey lines to estimate the thickness of the sediment layer. A non-tuned air gun array with a total capacity of 11 ℓ (700 in3) firing at intervals of 50 m was used as a controlled source for SCS data.
The P-wave velocity structures were obtained through the method reported in Kasahara et al. (2007) and Nishizawa et al. (2007, 2009). At first, we introduced the sedimentary layer thickness constrained by the SCS data. Then, we used the tomographic inversion method “tomo2d” coded by Korenaga et al. (2000) and forward modeling with trial and error using 2-D ray tracing (Fujie et al., 2000; Kubota et al., 2005). The first arrivals from the crust (Pc) and the uppermost mantle (Pn) were used in the tomographic inversion analysis. The size of the horizontal grid interval in the inversion is 0.5 km, and the vertical spacing gradually increases with the depth following the relationship 0.05 + (0.01 * depth (km))1/2 km. The velocity distribution inside the lower crust is derived mainly from later refraction arrivals propagating through the lower crust (Pg) and from reflection phases from the Moho (PmP). The depth of the Moho was confirmed by the travel times of PmP. The P-wave velocity of the uppermost mantle was determined from the Pn travel times.
We performed checkerboard resolution tests for the velocity models in order to obtain the resolution. The results show high reproducibility of the checkerboard pattern down to about 15 km in depth. In addition, we also modeled the amplitude data by calculating two-dimensional synthetic seismograms using the finite difference method E3D (Larsen and Schultz, 1995).
3.1 Record sections and crustal structure of MTr6
The Pn apparent velocities are considerably high, ranging between 8.4 and 8.9 km/s, as shown in Fig. 2(a, b). At first, we assumed that the Pn velocity is 8.4–8.9 km/s just below the Moho, whose depth was well constrained by a number of clear PmP arrivals. However, the calculated Pn travel times were shorter than the ones observed in this model, and the offsets of 30–40 km where the Pn appears as a first arrival on the observed record sections are smaller than those in the observed data. Consequently, we carried out tomographic inversions for two initial models in which the crustal velocity structure down to the Moho is fixed and the uppermost mantle velocity is constant at 8.0 km/s and 8.6 km/s, respectively. The results for both models showed that the P-wave velocity distribution in the uppermost mantle is heterogeneous in both the horizontal and vertical directions and has slower areas at offsets of 70–90 km and 130–165 km within 5 km below the Moho.
Finally, we introduced these heterogeneous structures in the uppermost mantle into the velocity model and conducted forward modeling to confirm the propriety of the inversion results for constructing a final structural model. In the final model shown in Fig. 3(a, b), the value of Vp immediately below the Moho is 8.0 km/s through the profile and the depth at which Vp = 8.5–8.7 km/s fluctuates in the horizontal direction. The depths from the Moho to the areas where Vp = 8.5–8.7 km/s are 4 km at the offset of 70–90 km, 2 km at 130–165 km and about 0.7 km for the remaining regions. The velocity structures of the slower portions as inferred from the tomographic inversion cannot be constrained with high precision since their velocity and thickness are in a trade-off relationship. We estimated the velocity range for these areas as 8.0–8.5 km/s.
3.2 Record sections and crustal structure of MTr8
The record sections of MTr8-7 and MTr8-15 (Fig. 2(c, d)) show clear first arrivals up to an offset of approximately 100 km. A notable difference from the sections of MTr6 is the smaller Pn apparent velocity, which is in the range between 7.9 and 8.0 km/s.
The crustal velocity model for MTr8 is shown in Fig. 3(e, f). The Vp of the top sedimentary layer is 1.6–2.5 km/s, and the thickness is less than 0.5 km. The crust consists of an 1.6–2.5 km thick upper crust with Vp = 2.5–6.8 km/s and a 4.4–5.5 km lower crust with Vp = 6.8–7.1 km/s. The total thickness of the crust is about 6.8–7.4 km. These characteristics are almost the same as those for MTr6 (Fig. 3(a)), with the exception of the slightly lower velocity in the lower crust in the southwestern part of MTr8. The Vp in the uppermost mantle of MTr8 is 7.9 km/s, which is about 0.6–0.8 km/s lower than that for MTr6. Moreover, unlike MTr6, no heterogeneity was detected in the MTr8 uppermost mantle.
4.1 Oceanic crust and uppermost mantle structure
The crustal velocity structures are almost the same for MTr6 and MTr8 in the Northwest Pacific Basin. The crust is composed of an upper crust with a high velocity gradient and a lower crust with a small velocity gradient. The total thickness of the crust is 6.2–7.4 km.
White et al. (1992) compiled a number of seismic profiles from previous studies conducted in the Pacific and Atlantic oceans. They reported that the average crustal structure composed of an upper crust with Vp = 2.5–6.6 km/s and a thickness of 2.11±0.55 km and a lower crust with Vp = 6.6–7.6 km/s and a thickness of 4.97±0.90 km. Our results are consistent with these characteristics. In addition, the total crustal thickness of 6.2–7.4 km as obtained in the 154–160 Ma area supports their suggestion that the crustal thickness appears to increase slightly with age, which is based on extremely few datasets from the older Pacific crust.
The uppermost mantle structure along MTr6 is characterized by a large heterogeneity immediately under the Moho. However, no clear relationships between the heterogeneity and the seafloor topography, the magnetic anomaly lineation, and the crustal structure were observed.
4.2 Velocity anisotropy in the uppermost mantle
The Pn velocity for MTr6 is around 8.5–8.7 km/s, which is significantly different from the value of 7.9 km/s obtained for MTr8, indicating the presence of velocity anisotropy of 7–10% in the uppermost mantle. MTr6 and MTr8 are nearly parallel and perpendicular to the magnetic lineation, respectively. In order to estimate the magnitude of anisotropy with high precision, additional survey lines running in different directions are necessary. Thus, the anisotropy of 7–10% as derived from only two lines in this study might be somewhat underestimated.
The result that the direction of higher velocity is perpendicular to the paleomagnetic lineation is consistent with anisotropy caused by the lattice preferred orientation of olivine crystals in the mantle when the oceanic plate was created at the mid-oceanic ridge (e.g., Hess, 1964; Raitt et al., 1969). The velocity range (7.9–8.7 km/s) and the velocity anisotropy correspond to those of harzburgite, which is the predominant component of the upper mantle at fast-spreading ridges (Ismail and Mainprice, 1998).
Shinohara et al. (2008) also detected anisotropy in the uppermost mantle velocity in the Northwest Pacific Basin. They estimated the anisotropy at approximately 5%, which is smaller than our result of 7%. One of the possible reasons for this difference might be the dependence on the spreading rate. In this regard, Gaherty et al. (2004) showed that anisotropy observed in a slowly spreading area in the Atlantic Ocean with a half-spreading rate of 0.8–2.0 cm/yr has a small magnitude of 3.4±0.3%. The half-spreading rates of the areas in our study and those in the northeastern area as presented in Shinohara et al. (2008) are estimated to be >8 cm/yr (Müller et al., 2008) and 4.83 cm/yr (Nakanishi et al., 1992) from magnetic anomalies, respectively. Although our results support this relationship, more data is necessary in order to confirm it.
In this study, we performed a precise estimation of an unusually high velocity of more than 8.5 km/s and of large anisotropy in the uppermost mantle in the old and intact area of the Northwest Pacific Basin, which could provide important constraints for the evolution of the oceanic lithosphere at a high seafloor spreading rate.
The authors are grateful to Prof. Emeritus J. Kasahara, Dr. R. Kubota, Mr. K. Kokai, and the members of Japan Continental Shelf Surveys Co. Ltd. for performing data processing and analyses, as well as to Dr. S. Kodaira and Dr. Y. Ohara for their helpful comments. We also thank the captains and the crew of S/V SHOYO and S/V TAKUYO, Japan Coast Guard, as well as the members of the Hydrographic and Oceanographic Department, JCG, for managing the seismic surveys. Most of the figures in this paper were prepared using the GMT graphic package developed by Wessel and Smith. Finally, the authors thank K. Michibayashi and D. Lizarralde for their critical and constructive review of this work, which helped to greatly improve our manuscript.