| We studied the electrical conductivity structure of the upper mantle and uppermost part of the lower mantle beneath the Pacific region using magnetic field observatory data and newly obtained electromagnetic field data from the OHP (see Section 2-5). Our findings indicate the importance of the effect of land-sea conductivity contrast in obtaining proper structures (Fujii and Utada, 2000). This implies that the electromagnetic induction problem is inherently a 3-D one related to the effect of seawater, even if a 1-D mantle is assumed. We developed computer codes that model global and semi-global electromagnetic induction in a 3-D medium based on an integral equation method; these models are utilized to obtain PREM-like 1-D reference conductivity structure beneath the Pacific (Utada et al., 2003) and for the first successful 3-D inversion of semi-global electrical conductivity structure in the mantle (Koyama, 2002; Fig. 2.3.1). |
 |
| It is well known that the toroidal magnetic field must exist in the Earthfs core in order for it to be a dynamo, although the component is confined in the core. Although the field cannot be detected by magnetic field observations at the surface of Earth, its signature is present in the geo-electric field and can possibly be detected by electric field observations using cables at the scale of thousands of kilometers. We began geo-electric field observations in the northwest Pacific with the aim of detecting the signature of the toroidal field (see Section 2-5). Long-term (decade-scale) variations in electric potential obtained to date indicate that decade-scale variation in the toroidal magnetic field at the core mantle boundary is 1?10 times that of the poloidal field variation, with reasonable electrical conductivity of the mantle (Shimizu et al., 1998). We confirmed that the amplitude is consistent with the electrodynamics of the geodynamo (Shimizu and Utada, 2004). |
| We are studying the small-scale magnetohydrodynamics within a rotating system as a step towards obtaining better parameterization of subgrid scale phenomena, especially the mean-field electromotive force, and establishing more realistic geodynamo models as a joint project with David E. Loper (Florida State University, USA) and Arnaud Chulliat (Institut de Physique du Globe de Paris, France). Structures of the flow and magnetic field driven by a buoyant blob are classified based on the strength of the rotation of the system and the magnetic field (Shimizu and Loper, 1997); detailed structures are obtained for the case of rapid rotation that is appropriate for the Earthfs core (Loper et al., 2003; Chulliat et al., 2004; Fig. 2.3.2). Although the integral of the electromotive force to the flow and field over entire space is zero, it has a significant non-zero value when integrated over the cross-section of the Taylor column, and it is significantly anisotropic (Shimizu and Loper, 2000; Chulliat et al., 2003, 2004). |
 |
| The problem of the state of the inner core boundary (ICB), whether a mushy layer or a slurry layer, is revisited as a new joint research project with J.-L. LeMouel and J.-P. Poirier (Institut de Physique du Globe de Paris, France). Stability analysis of the solid-liquid interface indicates that the condition that leads to a slurry layer does exist, but it cannot be attained within the Earthfs core; it is most probable that a mushy layer exists at the ICB. |
| Determining core surface flow by inverting geomagnetic secular variation with a frozen flux assumption has unavoidable physical non-uniqueness. We studied the use of length of the day (l.o.d.) data to overcome the non-uniqueness by assuming topographic (Asari et al., 2006) or electromagnetic core?mantle coupling. Cylindrical torque, which should be balanced in decade-scale core dynamics, is additionally considered both as a constraint to reduce the non-uniqueness and to obtain dynamically appropriate core surface flow (Asari, 2006). We found that only topographic core?mantle coupling is compatible with decadal cylindrical core dynamics. |
| The highlight of tomographic studies within the OHP is the introduction of a unified story for the fate of subducting slabs (Fukao et al., 2001; Fig. 2.1.2?3). After Fukao et al. (1992) suggested the existence of stagnant slabs, various tomographic studies attempted to reveal further detailed slab structures. Compiling those results, we found that every major subduction zone in the world can be explained as a snapshot of the following subduction process: subducting slabs tend to be deflected within the Bullen transition zone (400?1000 km depth), but can sink into the lower mantle via an instability. To confirm the plausibility of this idea, we are now conducting a nation-wide project to reveal the nature and role of stagnant slabs. |
| Tomographic studies conducted within the OHP have also been characterized by efforts to develop unique techniques. We developed an efficient method to compute accurate synthetic seismograms for 3-D heterogeneous Earth models (Takeuchi et al., 2000), an appropriate weighting method to correct heterogeneous data sampling (Takeuchi and Kobayashi, 2004), and accurate methods for travel-time measurements to correct systematic bias in conventional methods (Fukao et al., 2002; Oki et al., 2004). These unique techniques have revealed important features of the Earth structure, including (i) boundary layers at 670 km depth indicated by the predominance of lateral heterogeneities with longer horizontal scale lengths in this region, and (ii) low R (=dlnVs/dlnVp) values in subducting slabs obtained by highly accurate S?P time measurements (Fig. 2.3.3). |
 |
| Although seismic tomography is a powerful tool to unravel hidden structures within the Earth's deep interior, its resolution remains limited. To supplement existing information on the Earth's deep interior, we used the array seismological technique, which is sensitive to small-scale (~10 km) heterogeneities. We have particularly focused on utilizing dense and high-quality Japanese network data (J-array, F-net, Hi-net) in an attempt to detect structures that were not detected in previous studies. Our earlier studies indicated the existence of discontinuities/reflectors in the mid-mantle (Kawakatsu and Niu, 1994; Kawakatsu and Niu, 1997; Vinnik et al., 1998, 2001; Fig. 2.1.4) and multiple discontinuities at the bottom of the mantle transition zone (Niu and Kawakatsu, 1996); other researchers have investigated both of these features in subsequent studies. We also succeeded in estimating the local density contrast at discontinuities in the mantle transition zone, which is a difficult parameter to estimate seismologically (Kato and Kawakatsu, 2001). |
| (1) Reflection seismology of the upper mantle |
| Recent studies utilize the large amount of high-quality Hi-net data. For example, to delineate the effect of the presence of a cold "deep mantle slab" on the mantle transition zone discontinuities, we conducted a receiver function analysis to finely map the seismic discontinuities beneath the Japanese Islands (Kawakatsu and Watada, 2005; Fig. 2.3.4). The obtained image shows remarkably clear continuous features corresponding to the 410 km and 660 km transition zone discontinuities; in addition, the top surface of the subducting Pacific plate is traceable to at least 300 km depth. While these general features are expected for a penetrating subduction, the observed 500-km-wide depression of the 660 km discontinuity beneath southwest Japan appears to be consistent with neither a simple penetrating slab nor a flat stagnating slab near the boundary. A model that involves a more complicated morphology of the stagnating slab structure may be required. |
 |
| (2) Reflection seismology of the inner core |
| We have also made an unusual entire array observation of the near-vertical PKiKP phase, which is known to be very difficult to observe (Kawakatsu 2006, in press). Array analyses of this rare data set show a sharp (<~1 km) inner core boundary. Utilizing the PKiKP as a reference phase, a reflection seismological study of the inner core was possible for the first time. This study identified a possible discontinuity beneath the ICB. |
 |
 |