M. Hyodo et al., GEOMAGNETIC SECULAR VARIATION RECONSTRUCTED FROM MAGNETIZATIONS OF WIDE-DIAMETER CORES OF HOLOCENE SEDIMENTS IN JAPAN, Journal of Geomagnetism and Geoelectricity, 45(8), 1993, pp. 669-696
A record of the secular variation of the geomagnetic field for the las
t 1 1000 years has been obtained from the remanent magnetization of se
ven wide-diameter (20 cm) cores of marine and lacustrine sediment from
central to south-west Japan. Magnetization directions of five cores,
possessing high-amplitude variations, exhibit good correlation. Two co
res show low-amplitude variations, but they can be correlated well wit
h other records after deconvolution to remove filtering effects of the
post-depositional magnetization process. A composite secular variatio
n curve was constructed by stacking the field direction records. Time
constraints were obtained from radiocarbon ages of shell and wood frac
tions or tephrochronology. The secular variation curve agrees well wit
h the archeomagnetic record after 1400 yrBP, with a slight difference
in the time range 1400-2000 yrBP. The secular variation for the last 1
1000 years shows an elongated distribution of VGP's. The azimuth of th
e elongation, about 40-degrees-E, is consistent with that of VGP's fro
m Japanese volcanic and sedimentary rocks during Brunhes epoch. The an
gular standard deviation is 14.5-degrees (upper limit = 15.4-degrees,
lower limit = 13.7-degrees). This estimate is slightly larger than tha
t observed globally during Brunhes epoch and the position of the avera
ge VGP deviates 8.5-degrees from the geographic north pole. These VGP
analyses suggest a stationary nondipole source. The secular variation
records from five northern hemisphere sites between 135-degrees-E and
95-degrees-W in longitude possess a single prominent feature: an extre
me easterly swing in declination. A plot of age versus longitude of th
e swing at each site shows a clear westward drift, at a rate of about
0.13-degrees/yr. The field vector around the swing represents clockwis
e looping at all the sites. This suggests that the swing is caused by
a large non-dipole source which drifts at least from 135-degrees-E to
95-degrees-W in longitude.