VELOCITY ANISOTROPY IN THE SEA OF JAPAN AS REVEALED BY BIG EXPLOSIONS

Transcription

1 J. Phys. Earth, 26, Suppl., S 491-S 502, 1978 VELOCITY ANISOTROPY IN THE SEA OF JAPAN AS REVEALED BY BIG EXPLOSIONS Hiroshi OKADA,*1 Takeo MORIYA,*1 Toru MASUDA,*2 Takeshi HASEGAWA,*3 Shuzo ASANO,*4 Keiji KASAHARA,*5 Akira IKAMI,*6 Harumi AOKI,*6 Yoshimi SASAKI,*7 *1 Faculty of Science, Hokkaido University, Sapporo, Japan *2 Faculty of Science, Tohoku University, Sendai, Japan *3 Akita Technological College, Akita, Japan *4 Earthquake Research Institute, University of Tokyo, Tokyo, Japan *5 National Research Center for Disaster Prevention, Tokyo, Japan *6 Faculty of Science, Nagoya University, Nagoya, Japan *7 Faculty of Education, Gifu University, Gifu, Japan *8 Disaster Prevention Research Institute, Kyoto, Japan (Received June 19, 1978; Revised September 18, 1978) Seismic waves generated by two explosions of dynamite, 5 tons each, in the Sea of Japan off the coast of northern Honshu were observed at more than 100 temporary and permanent seismological stations in the Hokkaido, Honshu, Sado, and Oki islands. A purpose of these measurements was to extend our investigation of lateral variation in Pn velocity which has been found around northeastern Japan in the previous explosion experiments. In fact, a lateral variation in Pn velocity by about 5% was confirmed in regions of the uppermost mantle below the Sea of Japan and the Honshu island, although the boundary where the velocity change takes place was not determined. The measurements have also revealed an indication that the upper mantle just beneath the Moho interface under the area in the southeastern half of the Sea of Japan is anisotropic with respect to P-wave velocity. The velocity variation in the anisotropy is approximately 0.4km/sec (i.e., 5%) about a mean velocity of 7.94km/sec. The direction of the dicular to the general trend of northern Honshu as well as to magnetic lineations. 1. Introduction Until recently, seismic anisotropy of the uppermost mantle has been revealed in a number of oceanic areas (RAITT et al., 1969; MORRIS et al., 1969; KEEN and TRAMONTINI, 1970; KEEN and BARRETT, 1971) and has been considered as a typical and exclusive property of the upper mantle under the ocean. More recently, explosion-seismic experiments provided an evidence that the seismic anisotropy is also present in the upper mantle under the continent (BAMFORD, 1973, 1976, 1977). Furthermore, in the long-range observations of large explosions, HIRN (1977) presented some evidence of anisotropic propagation of mantle phase beneath the European continent, from which an anisotropic layer was inferred at depths where the lithosphere-asthenosphere transition is supposed to be. S 491

2 S 492 H. OKADA et al. RAITT et al., 1971) and also to be parallel to the tectonic trend in the continent (e.g., FUCHS, 1977). The Sea of Japan is a marginal sea with two major structural provinces, separated is predominant, including the Yamato Rise, Tsushima Basin, Yamato Basin, and the ridges and troughs along the continental slope off Honshu; the dominant structural trend of the ridges and troughs is roughly northeast to southwest. In the Sea of Japan, magnetic surveys have extensively been carried out by YASUI et al. (1967) and ISEZAKI et al. (1971). Using all the available geomagnetic data in the Sea of Japan, ISEZAKI and UYEDA (1973) established that sublinear magnetic anomalies run subparallel to the general trend of the Japanese islands. The Sea of Japan is also characterized with high heat flow which indicates that the isotherms are anomalously raised under the Japan Sea bottom (WATANABE, 1972). From all these features, reflecting the dynamic processes in the crust and upper mantle, it was inferred that some seismic anisotropy would be induced in the upper mantle under the Sea of Japan. In 1976 the Research Group for Explosion Seismology carried out explosion-seismic experiments with two shots in the Sea of Japan off northern Honshu and with observation stations in the Hokkaido, Honshu, Sado, and Oki islands. The main purpose of the experiments was to extend our investigation of the lateral variation in Pn velocity, which had been studied during the preceding few years in the regions from the Honshu island to the Pacific area around the Japan trench. The distribution of the shot points and observation stations in the experiments was suitable for the purpose, but of less ideal for measurements of the anisotropy in the uppermost mantle under the Sea of Japan. 2. Explosions and Observations Explosives, 5 tons of dynamite for each shot, were fired at a depth of 175m in the Sea of Japan about 250km off northern Honshu on July 28 and 30, The two explosions in the experiments are the fourth and fifth ones in a series of experiments conducted as a part of the Japanese Geodynamics Project, in which "Seiha Maru," a salvage ship of Nippon Salvage Company, was used as the shooting vessel. The shot points will then be referred to as SEIHA-4 and SEIHA-5, respectively. The locations of the shot points are respectively, which were fixed by satellite navigation. The shooting procedure was the same as that in the previous experiments (ASANO et al., 1978; OKADA et al., 1978). For the two shots Pn arrivals were recorded at 103 temporary and permanent seismological stations in the Hokkaido, Honshu, Sado, and Oki islands. In the experiments an ocean bottom seismometer was also anchored at a point about 160km east of shot point SEIHA-4, but because of anomalous arrival times the data obtained by the OBS were not used in the analysis. Figure 1 shows the locations of the shot points and observation stations. Crosses are the shot points and circles, the stations. In the figure, four profiles, P1, P3, P4, and P5 of the seismic refraction surveys by Murauchi et al. (JAPANESE NATIONAL COMMITTEE FOR THE UMP, SCIENCE COUNCIL. OF JAPAN, 1967) are also shown. The results from these surveys will be referred to later. Among these stations we selected 27 stations which are located less than 50km land-

3 Velocity Anisotropy in the Sea of Japan as Revealed by Big Explosions S 493 Fig. 1. Shot points and observation stations. Crosses designated by SEIHA-4 and SEIHA-5 represent the shot points; solid circles show the stations at which the data used in the analysis were obtained and open circles show other stations. Segments designated by P1, P3, P4, and P5 are the seismic refraction profiles by Murauchi et al. (JAPANESE NATIONAL COMMITTEE FOR THE UMP, SCIENCE COUNCIL OF JAPAN, 1967) wards from the coastline along the wave paths. The selection of the stations was mainly to focus the present study on data pertinent to the properties of the uppermost mantle beneath the sea. Such data are to be provided by the selected stations because these stations have 30-50km offset distance of Pn arrivals and more than 80% of the wave paths to these stations are under the Sea of Japan. The selected stations, represented by solid 3. Travel Time Data The Pn arrivals from the two shots were observed at a total of 44 stations: 18 for SEIHA-4 and 26 for SEIHA-5. Travel times of the arrivals were determined with an accuracy better than 0.1sec. Clearly the amount of data was so little that the crust and upper mantle structures could be derived neither beneath the shot points nor beneath the observation stations. For the present purpose, however, it is enough that only the Moho-time terms can be estimated at the shot points and stations. The Moho-time terms at the shot points could

4 S 494 H. OKADA et al. be estimated on the basis of the crust and upper mantle structures derived by Murauchi et al. (JAPANESE NATIONAL COMMITTEE FOR THE UMP, SCIENCE COUNCIL OF JAPAN, 1967) for four refraction profiles P1, P3, P4, and P5 (Fig. 1). The smallest one among the Mohotime terms given by the structures is 3.04sec at profile P3 located very near the northern Honshu, the largest one 3.81sec at profile P1 located in the Japan Basin, and the average Throughout the analysis, the average time term is assumed for the Moho-time terms at both shot points. The Moho-time terms at stations except ones where crustal structures were available were estimated by using the relationship between Moho-time term and Bouguer gravity anomaly obtained by OKADA et al. (1978). The relationship will give time terms in which errors are less than 0.33sec that has little effect on the results to be obtained. The Bouguer gravity anomalies required at stations were provided by the map of Bouguer gravity anomalies in Japan (TOMODA, 1974). If the Moho-time terms estimated for a shot point and station are subtracted from original travel time of the station, the residual gives a time of propagation taken by waves which propagate from a point under the shot point to a point under the station in the mantle just below the Moho interface. The residual time will be referred to as the mantle travel time for simplicity. Figure 2 shows the mantle travel times for the two shots plotted against distances with the reduction velocity 8.0km/sec. As shown in the figure, the mantle travel times are largely scattered between -2 and +2sec with an indication that most arrivals for SEIHA- 4 are early relative to those for SEIHA-5 at stations 300 to 500km distant from shot point. From this figure, it is difficult to find any dependence of travel times on distance. However, it should be noted that a velocity given by a line roughly fitting these travel time plots is 7.9km/sec which is higher by 5% than the velocity in the topmost mantle under northern Honshu (ASANO et al., 1978). The difference between these velocities should be taken as an evidence for a lateral change in the velocity in the upper mantle below the Sea of Japan and the Honshu island, although the boundary where the velocity change occurs is not determined. The mantle travel times are plotted again as a function of azimuth in the upper part of Fig. 3, the lower part of the figure being the corresponding distance. As can be seen in the figure, the mantle travel times for both shots are distributed so as to supplement with each other, and as a result the variation of the mantle travel times shows a clear dependence on the azimuth. It may be a question whether the observed azimuthal dependence of the mantle travel time is a spurious one resulting from the distribution of the stations with different distances. However, the question was to some extent answered by con-

5 Velocity Anisotropy in the Sea of Japan as Revealed by Big Explosions S 495 Fig. 3. Reduced mantle travel times and their corresponding distances versus azimuth. Fig. 4. Pn interval velocities as a function of azimuth. verting the mantle travel time into Pn interval velocity, with which the Pn phase traveled from a point under the shot point to a point under the station in the upper mantle just below the Moho interface. The Pn interval velocity is the velocity given by dividing the distance between shot point and station by the mantle travel time. Figure 4 shows the Pn interval velocities as a function of azimuth. The interval velocity also has a clear dependence on the azimuth. Figure 5 is a map to show spatial distribution of wave paths along which the interval velocity is either higher or lower than the average interval velocity. The stations in northern Honshu provide an evidence that the azimuthal dependence of the interval velocity is not produced by local effects of certain stations where arrivals are always either early or late. This evidence is also given in Table 1 in which listed are the

6 H. OKADA et al. Fig. 5. Map showing spatial distribution of wave paths along which the interval velocity is either higher (solid line) or lower (dotted line) than the average interval velocity 7.95km/sec. Table 1. Data on azimuths, distances, reduced mantle travel times, and interval velocities for two shots at stations FUT, OGR, NIB, IWM, OGA, and HOJ in northern Honshu. data on azimuths, distances, reduced mantle travel times, and interval velocities for the two shots at 6 stations: FUT, OGR, NIB, IWM, OGA, and HOJ in northern Honshu. Table 1 provides that the interval velocities for SEIHA-4 are higher than the average interval velocity 7.95km/sec and those for SEIHA-5 lower than the average. While the distance differences at a station for both shots are less than 20%, most differences in the reduced travel time between both shots evidently exceed the errors included in the Mohotime terms at shot point and station. From these data, it may be concluded that the dif-

7 Velocity Anisotropy in the Sea of Japan as Revealed by Big Explosions S 497 ference of interval velocities for both shots is not a spurious but actual one and will be attributable to the nature of the uppermost mantle under the area of investigation. 4. Results of Analysis Among various possibilities to cause the azimuthal dependence of the reduced travel times, or of the interval velocities are lateral variations in Pn velocity and velocity anisotropy in the uppermost mantle under the sea. In the present paper, we shall try to analyze the data on the assumption that the observed azimuthal variation is due to the velocity anisotropy existing in the uppermost mantle under the Sea of Japan. According to BACKUS (1965), for a small anisotropy the P-wave velocity Vp may be represented by a constant term and a perturbation term that depends upon azimuth, that is, responds to the Pn velocity observed and the azimuth is measured clockwise from the north. If such anisotropy is assumed in the uppermost mantle over which the crust has a gently undulating structure beneath the area of investigation, the theoretical Pn between the i-th shot point and the j-th station may be written as (2) where V0 is the velocity averaged over all angle and a's represent the Moho-time terms corresponding to this velocity of the i-th and j-th area, respectively. Dij is the distance between the shot point and station, and R's are the offset distances at i-th and j-th area, respectively. The method to measure the anisotropy based on a set of travel times, Eq. (2), has been developed by RAITT et al. (1969) and MORRIS et al. (1969). The method is particularly suited to experiments in which large amounts of data are available from uniformly distributed shot points and stations. If least squares analysis is applied to the data, not only the numerical values of A, B, C, and D of Eq. (1), or of Eq. (2), that is, the variation of velocity with azimuth but also the time terms and the averaged mantle velocity will be determined. In the present experiments, the observations of Pn arrivals were made on a semicircle; furthermore, the amount of data is relatively small so that it is difficult to calculate even the time terms. Therefore, direct application of the method to the data seems to be impractical. In the present analysis, however, not only the Moho-time terms but also the offset distances at the shot points and stations may be taken as known quantities. Of course, the Moho-time terms and the offset distances include errors due to the fact that these quantities have been estimated without taking account of anisotropy in the mantle. Practically the uncertainties were too small to produce serious errors in the solutions. were ignored since the coefficients C and D are in general much smaller than the coefficients A and B, and the offset distances at all the calculated from the structures. The structures were determined by the past seismic refraction experiments in the Japanese islands. In addition, the offset distances at both shot (1)

8 S 498 Fig. 6. Velocity anisotropy given as deviations from mean velocity of 7.94km/sec plotted against azimuth. Enclosed crosses are the values derived from seismic refraction profiles (JAPANESE NATIONAL COMMITTEE FOR THE UMP, SCIENCE COUNCIL OF JAPAN, 1967). Shaded portion represents the range within which a direction perpendicular to magnetic lineations (ISE- ZAKI and UYEDA, 1973) is obtained. Vertical bars represent errors in velocity which result in ambiguity in the Moho-time terms and offset distances at shot points and observation stations. estimated from the structures in profiles P1, P4, and P5. Figure 6 shows a least-squares fit of the velocity anisotropy given as deviations from the mean velocity. In the figure, the refraction results obtained by Murauchi et al. (JAPANESE NATIONAL COMMITTEE FOR THE UMP, SCIENCE COUNCIL OF JAPAN, 1967) in the Sea of Japan are also shown. The refraction results at each profile is plotted on two azimuths direction of their results. As can be seen in the figure, the refraction results except for the Japan Basin well supplement the results we obtained. As the Japan Basin has topographic features different from those in the area of the present investigation (HILDE and WAGEMAN, 1973), the directions of anisotropy may be also different in the two areas. The maximum velocity obtained occurs in a southeast-northwest direction, that is, The difference between the maximum and minimum velocities amounts to about 5%, which is within the range of anisotropy in the uppermost mantle observed to date in the Pacific Ocean. The mean velocity was obtained to be 7.94km/sec which is lower 4% than those observed in the Pacific Ocean. Table 2 gives the least squares regression results for an anisotropic upper mantle as well as an isotropic ones. Comparison between the results shows that standard error about the regression for anisotropy is smaller than that for isotropy, and hence allows us to conclude that the anisotropic upper mantle under the Sea of Japan is probably real.

9 Velocity Anisotropy in the Sea of Japan as Revealed by Big Explosions S 499 Table 2. Comparison of least-squares regression results of anisotropic upper mantle with those of isotropic one. 5. Discussion For the possibilities to cause the azimuthal dependence of the reduced travel times, or of the interval velocities, the lateral variations in Pn velocity and velocity anisotropy may be considered. The lateral variations in Pn velocity may be produced by such a model that a low velocity material occurs in several regions in the upper mantle. In this model, the low velocity material has to be located at both regions off southwest Honshu and off southwest Hokkaido. A sequence of thin vertical layers with alternating high and low velocities such as dyke injections is a model of the layering anisotropy (KUMAZAWA, 1964; FUCHS, 1977). Based on the model, the dyke has to be oriented perpendicular to the general trend of the Japanese islands, or unlikely large fraction of the dyke material has to be located in the uppermost mantle under the Sea of Japan. However, no evidences to suggest these models have been reported. At the present stage, these models are unlikely to be applicable to our case. The most favorable model to explain the observed azimuthal dependence of the reduced travel times, or of the Pn interval velocities is the velocity anisotropy due to crystal anisotropy in the upper mantle (DE ROEVER, 1961; HESS, 1964). Although pyroxenes in eclogites show considerable preferred orientation, eclogites in bulk do not show significant anisotropy (KUMAZAWA et al., 1971). On the other hand, olivine grains in peridotitic rocks usually show the pronounced preferred lattice orientation (PHILLIPS, 1938; TURNER, 1942) and the presence of large velocity anisotropy in peridotitic rocks has been confirmed by a number of authors (e.g., CHRISTENSEN, 1966; KASAHARA et al., 1968; KERN, 1978). If the genetical and evolutional environments (temperature, stress and the resulting deformation) in the upper mantle are uniform in a region, the pattern of preferential orientation of olivine is expected to be uniform, giving rise to the uniform anisotropy in the region. On the assumption of uniformity in the anisotropy type as well as in the mean velocity in the upper mantle under the Sea of Japan, we have obtained the anisotropy of 5% and the mean Pn velocity of 7.94km/sec. The degree of anisotropy 5% obtained here is quite reasonable in magnitude for peridotitic rocks and is also the same magnitude as those observed in the upper mantle under the Pacific Ocean (RAITT et al., 1969; MORRIS et al., 1969). However, the mean velocity 7.94km/sec is lower by 2-4% than those reported in the Pacific Ocean. This difference raises a question whether the present peridotite model for the upper mantle under the Sea of Japan is adequate or not. To account for the low mean velocity observed, we have referred to the measurements of velocities of peridotite under the simultaneous action of high temperature and high confining pressure by MATSUSHIMA and AKENI (1977), and KERN (1978). According to their measurements, the low mean velocity may be observed as a velocity of the peridotitic

10 S 500 H. OKADA et al. upper mantle only if temperature at depths in the upper mantle under the Sea of Japan In fact, heat flow measured in the Sea of Japan is anomalously high (YASUI et al., 1968) as indicating that the isotherms are abnormally raised under the sea bottom. Based on the heat flow measured, the temperature difference under the Sea of Japan and the (WATANABE, 1972), which would have yielded the observed low mean velocity. The direction of the maximum velocity seems to roughly agree with a direction perpendicular to the axis of the northern part of Honshu as well as to the long axes of three main ridges in the Sea of Japan: the Yamato, Oki, and Sado ridges, which are located in the area of investigation. In Fig. 6, the shaded portion shows the range of azimuth perpendicular to magnetic lineations (ISEZAKT and UYEDA, 1973). As can be seen in the figure, a surprisingly close agreement between the direction of the maximum velocity and that perpendicular to the magnetic lineations can be recognized. Such agreement has been observed also in the northeast Pacific where significant anisotropy has been observed (RAITT et al., 1969; MORRIS et al., 1969). The present paper is the first to suggest the presence of velocity anisotropy in the uppermost mantle under the Sea of Japan, the marginal sea. Although more detailed measurements might be required to confirm the suggested anisotropy, the result we obtained would be an evidence in favor of the view that the Japan Sea floor evolved through a similar spreading process to that occurring at the midoceanic ridges. 6. Conclusions Seismic waves generated by two explosions, 5 tons each, in the Sea of Japan off the coast of northern Honshu were observed at more than 100 seismic stations in the Japanese islands. First arrival data obtained at 27 selected stations which are located on the Japan Sea side of Hokkaido and Honshu islands as well as in Sado and Oki islands were analyzed. Although the observations were made only on a semicircular area, the results presented here demonstrate that the anisotropy of Pn velocity is required by the data. The velocity variation is approximately 0.4km/sec (i.e., 5%) about a mean velocity of 7.94km/sec, The Nippon Salvage Co., Ltd. gave us the opportunity to realize big explosions at sea by the salvage ship, Seiha Maru. Mr. K. Kobayashi, Captain of Seiha Maru, and all her officers and crew provided invaluable cooperation in explosion work. Our hearty thanks are also due to Nippon Yushi Co., Ltd. and Okanishi- Shibuya Maito Co., Ltd. for their cooperation in the explosion work. We thank the members of the micro-earthquake observatories of universities, who placed their data at our disposal, and the members of the Research Group for Explosion Seismology, who cooperated in the observation program. We would also like to thank Professor M. Kumazawa of Nagoya University for critically reading the manuscript. Financial aid was granted by the Japanese Geodynamics Project and Earthquake Research Institute. A part of the numerical computation was carried out by FACOM at the Hokkaido University Computing Center (Problem No. 1001FO0401).

11 Velocity Anisotropy in the Sea of Japan as Revealed by Big Explosions S 501 REFERENCES ASANO, S., H. OKADA, T. YOSHII, K. YAMAMOTO, T. HASEGAWA, K. ITO, S. SUZUKI, A. IKAMI, and K. HAMADA, Crust and upper mantle structure beneath Northeastern Honshu, Japan as derived from explosion seismic observations, submitted to J. Phys. Earth, BACKUS, G.E., Possible forms of seismic anisotropy of the uppermost mantle under oceans, J. Geophys. Res., 70, , BAMFORD, D., Refraction data in Western Germany-A time-term interpretation, Z. Geophys., 39, , BAMFORD, D., MOZAIC time-term analysis, Geophys. J.R. Astr. Soc., 44, , BAMFORD, D., Pn velocity anisotropy in a continental upper mantle, Geophys. J.R. Astr. Soc., 49, 29-48, CHRISTENSEN, N.I., Elasticity of ultrabasic rocks, J. Geophys. Res., 71, , DE ROEVER, W.P., Mantelgesteine und Magmen tiefer Herkunft, Fortschr. Mineral., 39, , mantle, Geophys. J.R. Astr. Soc., 49, , HESS, H., Seismic anisotropy of the uppermost mantle under oceans, Nature, 203, , HILDE, T.W.C. and J.M. WAGEMAN, Structure and origin of the Japan Sea, in The Western Pacific Island Arcs Marginal Seas Geochemistry, edited by P.J. Coleman, pp , University of Western Australia Press, Nedlands, Western Australia, HIRN, A., Anisotropy in the continental upper mantle: Possible evidence from explosion seismology, Geophys. J.R. Astr. Soc., 49, 49-58, ISEZAKI, N. and S. UYEDA, Geomagnetic anomaly pattern of the Japan Sea, Mar. Geophys. Res., 2, 51-59, ISEZAKI, N., K. HATA, and S. UYEDA, Magnetic survey of the Japan Sea (Part 1), Bull. Earthq. Res, Inst., 49, 77-83, JAPANESE NATIONAL COMMITTEE FOR THE UMP, SCIENCE COUNCIL OF JAPAN, Geophysical and geological data in and around the Japan Arc, in Japanese National Report for the UMP, compiled by T. Rikitake, pp. 2-25, the National Committee for the UMP, Science Council of Japan, KASAHARA, J., I. SUZUKI, M. KUMAZAWA, Y. KOBAYASHI, and K. IIDA, Anisotropism of P-wave in dunite, J. Seismol. Soc. Jpn., 21, , 1968 (in Japanese). KEEN, C. and C. TRAMONTINI, A seismic refraction survey on the Mid-Atlantic Ridge, Geophys. J.R. Astr. Soc., 20, , KEEN, C.E. and D.L. BARRETT, A measurement of seismic anisotropy in the Northeast Pacific, Can. J. Earth Sci., 8, , KERN, H., The effect of high temperature and high confining pressure on compressional wave velocities in quartz-bearing and quartz-free igneous and metamorphic rocks, Tectonophysics, 44, , KUMAZAWA, M., The elastic constants of rocks in terms of elastic constants of constituent mineral grains, petrofabric and interface structures, J. Earth Sci. Nagoya Univ., 12, , KUMAZAWA, M., H. HELMSTAEDT, and K. MASAKI, Elastic properties of eclogite xenoliths from diatremes of the East Colorado Plateau and their implication to the upper mantle structure, J. Geophys. Res., 76, , MATSUSHIMA, S. and K. AKENI, Elastic wave velocities in the Ichinome-gata ultramafic nodules: Composition of the uppermost mantle, in High-Pressure Research Applications in Geophysics, edited by M.H. Manghnani and S. Akimoto, pp , Academic Press, New York, MORRIS, G.B., R.W. RAITT, and G.G. SHOR, Jr., Velocity anisotropy and delay-time maps of the mantle near Hawaii, J. Geophys. Res., 74, , OKADA, H., S. ASANO, T. YOSHII, A. IKAMI, S. SUZUKI, T. HASEGAWA, K. YAMAMOTO, K. ITO, and K. HAMADA, Regionality of the upper mantle around northeastern Japan as revealed by big explosions at sea. I. SEIHA-1 explosion experiment, submitted to J. Phys. Earth, PHILLIPS, F.C., Mineral orientation in some olivine-rich rocks from Rum and Skye, Geol. Mag., 75, , RAITT, R.W., G.G. SHOR, Jr., T.J.G. FRANCIS, and G.B. MORRIS, Anisotropy of the Pacific upper mantle, J. Geophys. Res., 74, , RAITT, R.W., G.G. SHOR, Jr., G.B. MORRIS, and H.K. KIRK, Mantle anisotropy in the Pacific Ocean, Tectonophysics, 12, , TOMODA, Y., Maps of Free and Bouguer Gravity Anomalies in and around Japan, University of Tokyo Press, Tokyo, TURNER, F.J., Preferred orientation of olivine crystals in peridotites, with special reference to New Zealand examples, Trans. R. Soc. N.Z., 72, , 1942.

12 S 502 H. OKADA et al. WATANABE, T., Heat flow through the sea floor, in Kaitei Butsuri, edited by M. Iwashita, Y. Komaki, M. Hoshino, S. Horibe, and J. Masuzawa, pp , Tokai University Press, Tokyo, 1972 (in Japanese). YASUI, M., Y. HASHIMOTO, and S. UYEDA, Geomagnetic studies of the Japan Sea-Anomaly pattern in the Japan Sea, Oceanogr. Mag., 19, , YASUI, M., T. KISHII, T. WATANABE, and S. UYEDA, Heat flow in the Sea of Japan, in The Crust and Upper Mantle of the Pacific Areas, Geophys. Monogr. Ser., No. 12, edited by L. Knopoff, C.L. Drake, and P.J. Hart, pp. 3-16, Am. Geophys. Union, Washington, D.C., 1968.