第 四 紀 研 究(The Quaternary Research) 38 (3) p. 202-208 June1999 Recent Progress in Paleomagnetic and Rock-magnetic Studies of the Quaternary in Japan Masayuki Hyodo* Detailed secular variation records of geomagnetic field direction during the Holocene, derived from archeomagnetic and sedimentary magnetic studies in Japan, can be used as a time scale for intervals of 102-103 years. Accumulated pale- ointensity data for the last 50kyr, from baked earth and volcanic rocks using the Thellier method, provide reliable age constraints for time intervals of 103-104 years. The geomagnetic excursions and short reversal events found in Japan, which mostly coincided with paleointensity minima at intervals of 104-105 years in the Pacific, are also useful for dating. Besides its chronological applications, magnetic analyses of sediments are useful in other areas of Quaternary studies. Key Words: paleomagnetism, rock magnetism, secular variation, excursion, short reversal event ternary chronological studies. Results of Qua- I. Introduction ternary magnetic studies in Asia are also The use of paleomagnetism in Quaternary reviewed. studies arose from the discovery of reversed II. Secular variation polarity fields during the early Pleistocene in Japan and Korea (Matuyama, 1929). This dis- 1. Field direction covery about the timing of field reversals even- There are observations and historical tually led to the concept of the geomagnetic records of geomagnetic field direction, but polarity time scale (GPTS) (Cox,1973), which is these go back only a few hundred years. A now an important dating tool. In addition to standard secular variation (SV) curve for geo- polarity reversals, geomagnetic secular varia- magnetic field direction in SW Japan over the tion (SV), excursions, and short reversal events last 2,000 years has been established from ar- recently have been incorporated into time cheomagnetic data (Hirooka, 1971). This SV scales. curve has been reinforced by subsequent ar- Since the pioneering work of Matuyama cheomagnetic studies (e. g. Maenaka, 1990). (1929) and Watanabe (1958), studies of past geo- Sedimentary magnetic studies have extend- magnetic fields have been actively carried out ed the time span of the SV record in Japan. A in Japan using baked earth, lake and marine composite SV curve for the last 11.5 ka in cen- sediments, cave deposits, and volcanic rocks. tral to SW Japan was constructed by stacking In the present paper, results of such studies by the SV data of five cores from the Inland Sea, members of a research group of specialists in Lake Yogo, and Lake Kizaki (Hyodo et al., 1993 rock magnetism and paleomagnetism in Japan, a). The data used in that study were obtained published mainly in the 1990s, are reviewed by using a wide-diameter (20cm) core sampler. from the viewpoint of their application to Qua- Besides this, SV records were obtained from Received January 30, 1999. Accepted March 7, 1999. * Research Center for Inland Seas & Dept . of Earth and Planetary Sciences, Kobe University. 1-1 Rokkodai, Nada-ku, Kobe, 657-8501, Japan. 1999年6月 Paleomagnetic and Rock-magnetic Studies 203 Fig. 1 Secular variation of declination (top) and inclination (bottom) in central to SW Japan The open circles with 95 per cent confidence limits show the composite data from sediments (Hyodo et al., 1993), and the open triangles show the archeomagnetic data (Hirooka, 1971, 1983; Maenaka, 1990). sediment cores in Beppu Bay, Kyushu (Ohno et studies, even though dating cave deposits is al., 1991), and in Lake Biwa (Ali et al., 1999). quite difficult. At present, SVs recorded in These SV records generally show similar fea- such deposits are used for dating and estimat- tures, except for the long-term trend of wester- ing the growth rates of speleothems (Morinaga ly declination before about 7ka observed in the et al., 1993, 1994). SV record by Hyodo et al. (1993a). The ab- 2. Field intensity sence of this feature in the other two records is The intensity of past geomagnetic fields has probably due to detrending in declination logs become important in geomagnetochronology that were strongly affected by core-twisting, as the volume of paleointensity data has in- which may have been caused by sampling with creased. In Japan, the accumulated data ob- a small-diameter (8cm) corer. The SV curves tained using the Thellier method from baked from archeomagnetic and sedimentary mag- earth at archeological sites (Sakai and Hirooka, netic studies are consistent for the past 2,000 1986) and from volcanic rocks (Tanaka, 1990) years as shown in Figure 1. An SV record in show a sinusoidal variation with peaks at sediment cores from Erhai Lake, southwest about 2,000 and 9,000 years ago. Since such a China, shows that many features of declination variation pattern with a period of 7,000 years and inclination seen in Japan also occur in has been observed in both Europe and Japan, it China (Hyodo et al., 1999). may reflect the nature of the dipole moment Morinaga et al. (1989) showed that stalag- (Tanaka, 1990). mites from different limestone caves in SW Tanaka et al. (1994) obtained paleointensities Japan have consistent paleomagnetic SV at 10-22 ka from volcanic rocks in Japan and records. This suggests a new method for SV New Zealand using the Thellier method, and 204 Masayuki Hyodo June 1999 showed that the data for the 14-22ka interval using the magnetization of deep-sea sediment agree with the broad paleointensity minimum cores in the Pacific. These workers showed for the period 10-50ka that was suggested by that prominent intensity drops occurred at McElhinny and Senanayake (1982). The virtu- about 30-50, 110 -120,180-200,280-300,380- al dipole moments for 0-50ka (Tanaka et al., 410 and 520-550ka in many regions of the 1994) show a dominance of low values about Pacific. These characteristic features can be half that of the present one prior to 15ka, with used as reliable datum levels in the Brunhes a minimum of about 1/8 of present values at chronozone. about 40ka. These characteristic features of III. Excursions and short events paleointensities can be used as reliable age con- straints. Hyodo and Minemoto (1996) revised the ages Yamazaki and Ioka (1994) measured consis- of the short reversal events found in the 200-m tent relative paleointensity changes in five Lake Biwa core (Kawai et al., 1972), based on hemipelagic clay cores from the West Caroline tephrostratigraphy (Machida et al., 1991). A Basin, western equatorial Pacific, and con- mid-age for the Biwa I event was dated at 116 structed a composite paleointensity record for ka, although it had originally been dated at the last 200 kyr, dated with oxygen isotope about 180 ka (Kawai et al., 1972). The Biwa I ratios. It shows prominent intensity drops at event is therefore correlated with the Blake about 40, 110 and 190ka, which correlate with event that was dated at 111-117ka (Zhu et al., the Lashamp excursion, and the Blake and the 1994). The short event, formerly known as the Biwa I events, respectively. The paleointensity Blake event (Kawai et al., 1972), is new. We record was further extended by Yamazaki et herein tentatively name it the Biwa 0 event. A al. (1995) to the Brunhes/Matuyama boundary mid-age for the Biwa II event was determined to be about 190ka. The Biwa III event, absent in the revision of Hyodo and Minemoto (1996), occurred at 230-250ka, since it lies close to the Ata-Th tephra, the level of which was recently determined (Kuwae et al., 1997). An intermedi- ate magnetic polarity direction dated at 0.22 Ma in the central Taupo Volcanic Zone, New Zea- land (Tanaka et al., 1996), and a reversal rec- orded in a stalagmite younger than 350 ka in western Japan (Morinaga et al., 1992), may cor- relate with one of the short events at Lake Biwa. Three additional excursions, named A, B, and C (Hyodo and Minemoto, 1996), were identified in the Lake Biwa data. Excursion A is just below the AT tephra. Ohno et al. (1993) found an excursion just below the AT tephra in a deep-sea sediment core from off Shikoku Island in the Pacific, and correlated it with excursion Fig. 2 A depth-age plot for the 200-m Lake Biwa A. Another excursion found below it in the core same core was correlated with B and/or C. The diamonds show radiocarbon dates, and the open These excursions may correlate with the Mono circles show the K-Ah (7.1ka), Oki (10.2ka), AT (28ka), Lake Excursion (Levi and Karlin, 1989), the Aso-4 (89ka), Ata (105ka), Aso-3 (123ka), Ata-Th (240 Auckland excursion from basalt lavas dated at ka, a mid-age of 230-250ka), Aso-1 (266ka), and Handa 25-50ka in New Zealand (Shibuya et al., 1992), (290ka) tephras (Matsumoto et al., 1991; Machida and Arai, 1992). Ages younger than the AT are in calendar and an event in marine isotope stage 2 present years. in a core at ODP site 650 A (Channell and Torii, 1999年6月 Paleomagnetic and Rock-magnetic Studies 205 1990). dealt with the identification of volcanic depos- A long core from Osaka Bay contains two its in regions 500-1000km apart in central and short reversal events with excursional fields southwestern Japan. Hayashida et al. (1996) (Biswas et al., 1999). One reversal is at 0.69 Ma demonstrated a good agreement in paleo- in the early Brunhes chron, and is named the magnetic directions between the Imaichi "Stage 17 event" , because it occurred during a pyroclastic-flow deposits in central Kyushu, period of eustatic high sea level during marine the Azuki tephra in the Kinki district, and the oxygen isotope stage 17.